Roles and Expectations#

One other important way that we do differently than a lot of companies is the way we assign responsibility. Instead of titles, which are linked to the concept of hierarchies and ranking, we use roles, and all people in the organization share a set of expectations.

What is expected from everyone#

The general expectations for anyone working at OpenCraft:

  • I can work from wherever I want, provided I have access to a reliable high-bandwidth internet connection good enough for video calls.
  • I can set my own work schedule, and work when I want during the week, as long as I remain available to answer to others (please see: the Communication section below)
  • (In general) I am not expected to work weekends (unless I'm behind on my commitments), which should remain an exception, not the rule! If you find yourself forced to work on week ends more than once per month, that likely reflects an issue with your time management that needs fixing. If I work week ends as a personal choice, I will not expect other team members to also work week ends.
  • I will make sure my cell's responsibilities are continuously being properly taken care of, by reviewing their status at least once per week.
  • I will work at least the amount of hours per week that is specified in my contract (e.g. 40 hours/week), averaged over each month, excluding scheduled holiday time.
  • I will ensure that clients and people on the cell that I'll be working with (e.g. code reviews) know my availability. (e.g. If I'm taking Friday off, I'll make sure people who need me to do code review know that).
  • When taking time off, I will follow the procedure described on the vacation's section.
  • If I'm unexpectedly sick or unavailable due to an emergency, I'll make every effort to notify my cell and ask the sprint firefighters to take over my duties.
  • I will provide my own hardware (laptop).
  • I will keep my computer(s) secure and up to date as described in the security policy.
  • I will record all hours that I spend on a task using the Tempo timesheets tool in JIRA (including things like setting up devstacks and time spent learning, which we all need to do and is an important part of the task, especially at the beginning, or when joining a project).
  • If I use another tool for tracking my time, I will update the JIRA Tempo timesheets within 24 hours (excluding weekend days).
  • I will attend the Open edX Conference every year, unless exceptional and important personal circumstances prevent me from being present (Note that team members who joined OpenCraft before July 31st 2018 are only encouraged to attend the conference, but not strictly required to).
  • As a conference attendee, I will submit a talk proposal and present it. If my talk isn't accepted, I will co-present a talk with someone else from OpenCraft.


  • If any client email is addressed to me, I will respond to it within 24 hours (excluding weekend days), or ensure that someone else on my cell does.
    • I will CC or forward to all client emails and replies, so that others can take over the thread with that client if needed, or refer to it for context.
    • If I can't answer immediately, I will at least send a quick reply to let them know we're working on a response, and when they should expect it.
  • I will reply to pings/emails from other OpenCraft team members within 24 hours (again excluding weekend days and scheduled time off).
    • But this should be a "worst case" scenario - completing the work on time is still the primary goal, so when someone is blocked and pings me for help, I will try to do as much as I can to unblock them quickly, rather than starting a 24h ping-pong cycle that takes up all the time in the sprint without accomplishing any work.
  • I will respond to pings on GitHub or GitLab within 24 hours (with the usual exclusion for weekend days and scheduled time off).
    • If a ping has no corresponding ticket, or the ticket is not scheduled for the current sprint, I will respond to the ping with an estimate for when they can expect a full response.
  • I will read and respond to forum threads:
    • On the announcements forum: within 24h, excluding week-ends & time off;
    • On the rest of the forums, within 2 weeks, except for the off topic forum, which doesn't need to be read or replied to at all.
  • I will be professional and polite when communicating with clients.
  • I will prefer asynchronous over synchronous processes (keep meetings to a minimum). A chat conversation is a form of a meeting.
    • Generally, discussions should happen first asynchronously on the JIRA tickets; if there is really something that can't be efficiently sorted out asynchronously, have a chat or schedule a meeting. JIRA might have long response cycles (around 1 day turnaround). If this time isn't enough to unblock someone and finish their sprint commitments, use Mattermost, even though it's more disruptive to people's workflow.
    • If you do have a synchronous conversation with someone about a particular task, post a summary of the result/decision from that conversation on the JIRA ticket for easier future reference.
    • When scheduling meetings, give them a precise agenda. For people using Calendly, like Braden and Xavier, book meetings there, as it allows us to avoid the scheduling overhead.
    • Try, as much as possible, to use a similar approach with clients - don't let them invade your days with meetings. Calendly is good for this too, as it allows to define time slots where you'll have the meetings, to minimize the disruption they cause to your day and productivity. If you want a calendly account, let Xavier know and he will set you up with the OpenCraft account.
  • I will post on public channels on Mattermost rather than private 1-to-1 channels whenever possible, even if the message is just for one person. This allows us to know what others are working on, and helps to replace the overheard discussions in physical offices - it can also be an occasion for someone else with knowledge about your issue to get the context, and to intervene if it is useful to the conversation.
  • I will make sure I communicate with my reviewer(s) on tasks about availability and timezone overlap if I didn't have knowledge about it before. I will use the contact sheet where necessary.
  • I will join the weekly sprint meeting of my cell on Mondays, unless I have scheduled that day off in advance.

Sprint tasks#

  • I will follow the process and expectations outlined in the pull request process section.
  • I will never add my code (even DevOps code!) to a production branch directly - I will always create a Pull Request.
  • I will always ensure my Pull Requests are reviewed by another OpenCraft team member before merging, except if I am a core team member and I'm merging trivial changes. In this exceptional case, I may merge the trivial changes before receiving a review, but I will then ensure all of those trivial changes are reviewed and acknowledged post-merge or post-deployment by another OpenCraft team member. Trivial changes include:
    • Small Open edX environment tweaks (example)
    • Minor/bugfix package version bumps (example)
  • I will only commit to work each sprint that I believe I can comfortably complete within the allotted sprint time (two weeks). Here, "complete" means "get to external review or get merged, deployed, and delivered."
  • As a core team member, I will avoid taking newcomer-friendly tasks unless they are urgent and there are no newcomers available to take them on. I will take on reviews of newcomer-friendly tasks, and allow time to provide mentoring and extra guidance.
  • I will get all of my tasks ready for internal review (by someone else on the OpenCraft team) by the end of Wednesday in the second week of the sprint at the latest. This will ensure that there is time for the review to take place, and for me to address the review comments and get the internal reviewer's approval in time. If a ticket potentially requires multiple review cycles, get it into review as early as possible. Schedule reviews with your reviewer to make sure they have time when you get your work ready for review.
  • I will spend time to plan each task at the beginning of each sprint, including scheduling the review with the reviewer. Make sure you have an answer for:
    1. When do each step of the task need to be completed?
    2. Which days will you work on which task?
    3. When do individual parts need to be completed to be on time?
  • It's also important to keep some margin on the sprint, in case something doesn't go as expected.
  • I will get my tasks to at least External Review (or "Deployed & Delivered" if no external review is required) by one hour before the sprint planning meeting for the next sprint. As this is also dependent on the internal reviewer, I'll make sure she/he will be available around the time I finish.
  • If it is looking like I will have trouble meeting one of these deadlines, I will inform the sprint firefighters (and epic owner) as early as possible, so they can get me additional help, start planning for contingencies, and/or warn the client.
  • If necessary, I will work overtime to complete my tasks before the end of the sprint.
  • I will prioritise stories that spilled over from the previous sprint.
  • I will "work from the right", prioritizing responding to external reviews as the highest priority, then doing reviews / responding to questions from others on the team, and finally working on implementing my own "In Progress" tasks.
  • I will be helpful to team members, responding to questions, helping with debugging, providing advice, or even taking over tasks if I have time and they are overloaded. This has precedence over starting to work on new tasks, when finishing a sprint early.
  • Once a sprint, I will review all of my tasks that are in "Long External Review/Blocked" and if needed, I will ping the external reviewer/blocker or pull the task into an upcoming sprint.


Roles can be taken on by willing people, and people will usually take on multiple roles, changing over time, based on interest and availability. This allows for a much smoother progression of responsibilities than the ladder climbing game.

Note that responsibility is also uncorrelated with compensation and raises, which are given the whole core team at once, based on time spent working at the company and overall financial results. Compensation isn't a good source of motivation beyond a certain level, and this approach removes a lot of politics.

If there is any role you would like to take up, the best way is to say it publicly - for example in the forum. Then when opportunities arise others will remember it, and point you to tickets you can own. Keep in mind that role assignments are intended to be permanent, so keep that in mind before picking a role. Ideally you should work in a role for at least one year before you consider swapping/dropping it.

Changing Roles#

Once someone takes up a role in a call there generally no need to reassign it unless someone leaves OpenCraft. While an appointment to a role is generally permanent, no one should feel stuck in a role if they are no longer comfortable in it. As such it is desirable to openly discuss with other members of the cell on the forum, and perhaps in the sprint meeting when such a situation arises.

It is the responsibility of the current person with a role to find a replacement in their cell and help their replacement during a transition period while they are still getting comfortable with their new responsibilities.

Note that for roles like Client Owner or DevOps Specialist where the role involves a fair bit of specialised knowledge or context, great care should be taken to find a suitable replacement with a similar level of knowledge and context.

Types of OpenCraft members#

There are three types of members at OpenCraft:

  • Additional members:
    • Short-term members (temporary contractors), who have been hired for a specific task, scope or period of time. They are the most external members of the team.
    • Newcomers on probation (2 months, renewable).
  • Core team member - the new recruits who have been confirmed become core team members. They differ from the other types of members in that they tend to have more team-based responsibilities. For example, core team members are on a weekly rotation schedule where they often have to take on some additional roles, including Sprint Firefighter, being on Discovery Duty, and occasionally leading our weekly sprint kick-off meeting.

However, in all other regards, all types of developers are put to the same expectations -- no politics or special treatment between short-term developers, newcomers, and core team members.

The only acceptable exception is when providing extra mentoring on tasks for newcomers (or anyone known to not have much context in the task), which is expected and useful.

Cell member#

The coordination on each of these responsibilities is to be assigned, each one to an individual member of the cell as a recurring task. However, the way this responsibility is handled beyond what is described in this handbook is the responsibility of the cell as a whole, with all of its members being considered as weekly reviewers.

Each responsibility can be handled through multiple tasks assigned to multiple owners, as long as the overall responsibility for the coordination of each item is assigned to a single person:

Cell manager - Recruitment#

The recruitment manager role is responsible for organizing the selection of candidate hires to match the needs of the epic planning spreadsheet for the cell, as determined by the epic planning manager. This includes: Creating the newcomer onboarding epic and tasks; Finding a mentor for the newcomer, ahead of their start date; * Organizing the review and confirmation of core developers: Ensuring that all core members of the cell participate in the review, as well as one core member from each other cell. Compiling the feedback from individual reviews in a single non-nominative list.

Cell manager - Sprint management#

  • Each cell has its own sprint board and tickets, and is responsible for handling the sprint, as well as its preparation and refinement.
  • Every Monday before the sprint planning meeting, review the sprint planning:
    • Tasks from the previous sprint, waiting for 3rd party, and in the backlog;
    • Epics;
    • Rotations - double-check if any of the people assigned to a rotation in the upcoming sprint are off, and if so that a backup will be available (eg. that the second firefighter will be present, or that a third firefighter has been assigned on those days).
  • Assigning the rotations. How to assign the rotation hours within the cell are up to that cell, as long as the number of hours matches a % of the total hours of work in the cell. The hours can be adjusted slightly if more or less firefighting is required:
    • Firefighting budget: 7.5% (30h/week for 10 full times, split between two firefighters each sprint)
    • Discovery budget: 2.5% (10h/week for 10 full times) - The discovery budget and the discovery duty allocation are two different things. The weekly discovery duty allocation (5h/cell/week) is to be used for new tasks that pop-up in the course of a sprint - it is funded by the discovery budget. Any leftover discovery budget can be used for discovery tasks that are planned in advance.
  • Handling the delivery to the client and the verification of the tasks by the clients.
  • Update the sprint commitments spreadsheet for your cell during the sprint planning meeting, every week, using the Sprints. When the person handling the sprint management is also leading the sprint planning meeting, this responsibility can be delegated to the person taking notes that sprint.

Cell manager - Epic planning#

  • Understand the lifecycle of en epic:
    1. Most epics start with a discovery based on a client requirement. (For internal projects, the client is OpenCraft).
    2. An epic is created based on the corresponding discovery and starts in the "Prospect" column.
    3. The discovery and corresponding estimates are shared with the client, and the epic is moved to "Offer / Waiting on client".
    4. If the client accepts the work, the epic is moved to "Accepted".
    5. Once actual development starts, the epic is moved to "In Development".
    6. Once the client's requirements are met, and the client has access to the completed work, the epic can be moved to "Delivered to client / QA".
    7. The epic should be moved back from "Delivered to client/ QA" to "In Development" if the client requests additional work or modifications that need development.
    8. When all the work in the epic is complete (for instance if all upstream PRs have been reviewed and merged) the epic can be moved to "Done".
  • Recurring epics are generally not based on a project or discovery, but are used to track work for different cell roles.
  • Each cell has its own Epics board and epics, and is responsible for ensuring that the projects those epics represent are being properly delivered based on the above lifecycle.
  • Every sprint, during the first week, you should evaluate the changes in status of epics over the past sprint. This will involve ensuring that:
    • Each epic has an epic update.
    • Delivered epics are moved to "Delivered to client / QA".
    • Completed epics are moved to "Done".
    • Blocked epics are moved to "Offer / Waiting on client".
  • Generally, moving an epic from "Prospect" to "Offer / Waiting on client" or to "Accepted" isn't controversial. As described above, there are clear steps that trigger each of those transitions. In case of any doubt about the correct status for an epic leave a comment on the epic.
  • Every sprint, compare the Epics board for your cell (Bebop, Serenity) with the corresponding sheet(s) of the epic planning spreadsheet. Ensure delivery and bugfix deadlines for individual epics are on target. Comment directly on the spreadsheet.
  • When an epic is completed, make sure that it is correctly reflected in the "Time / Estimates" sheet of the Epic Planning spreadsheet.
  • Maintain a count of the amount of time required to complete the accepted budgets over the next months in the epic planning spreadsheet for your cell. This is used to inform recruitment needs.
  • Keep a look out for completed discoveries. If a discovery is generic and the estimates it produced can be reused for further discoveries and estimates down the line, add it to the Price list sheet of the epic planning spreadsheet.
  • Keep a look out for newcomers and core members joining or leaving the team, and add their details to the epic planning spreadsheet.
    • For newcomers this includes adding onboarding time to the onboarding section and specifying (or updating) their availability in the availability section. Only make these updates when the corresponding information is fully public, i.e., the newcomer should know whether they have been accepted or not by the time these updates are made.
    • For core team members leaving the team, make sure to reset their availability for the remaining months of the year.
  • Once per sprint, during the first week, post an epic update in the "Epic Planning - " epic with the following checklist:
h4. Epic planning update (Sprint ???)

* ( ) Make sure all epics have epic updates.
* ( ) If an epic's status has changed make sure it has the correct status on the Epics board. Most importantly:
** ( ) If an epic has been completed, move it to "Done".
** ( ) If an epic has become permanently blocked on the client, move it to "Offer / Waiting on client".
* ( ) Compare the Epics board with the epic planning spreadsheet.
* ( ) Add/Update details in the "Time / Estimates" sheet of the epic planning spreadsheet:
** ( ) For [new epics| = Epic AND status in (Backlog, Offer, Accepted, "In development")] (i.e., epics with a status of "Prospect", "Offer / Waiting on client", "Accepted", or "In development").
** ( ) For [completed epics| = Epic AND status in (Done, Archived)] (i.e., epics with a status of "Done" or "Archived").
** Note that you can filter the list of new/completed epics down to epics from your cell via the "Project" filter.
* ( ) For in-progress epics:
** ( ) Evaluate the amount of time required to complete the accepted budgets over the next months and update the epic planning spreadsheet.
** ( ) Ensure delivery and bugfix deadlines of individual epics are on target (or are being actively discussed on the epics). Comment directly on the epic planning spreadsheet, pinging epic owners as necessary.
** ( ) Ensure the projects from your cell's clients are being properly delivered using the epic management process.
* ( ) For completed discoveries, see if any estimates can be useful more broadly and add them to the "Price list" sheet of the epic planning spreadsheet.
* ( ) Check the calendar and/or the [Looking for mentors|] thread for people joining/leaving the team in the next couple months and update the epic planning spreadsheet as necessary:
** ( ) For people joining the team, add availability and onboarding hours for the coming months.
** ( ) For people leaving the team, update their availability for the coming months. If prompted by your cell's sprint manager, help find new owners for clients and epics belonging to the people leaving.
* ( ) Check the calendar for vacations coming up in the next couple months. If someone will be away for a total of 1 week (or longer), [post a comment mentioning their availability|] in the epic planning spreadsheet.
* ( ) Check the descriptions of the Hosted Sites Maintenance and Small Projects & Customizations epics (SE-1690, SE-1693) and update the list of clients for your cell, following the existing format.
** ( ) Add info about new clients that we recently on-boarded. (Make sure to skip clients that haven't moved past the prospect stage.)
** ( ) Remove info about clients that we no longer work with. (Make sure they have been off-boarded completely before doing this.)
* ( ) Adjust client budgets for sustainability dashboard as necessary.

h4. Notes


Cell manager - Sustainability of the cell#

  • Each cell is meant to be a sustainable entity by itself: its members are the closest to most of the work that impacts its sustainability: the successful estimation and delivery of each client project.
  • Some of the budgets for internal/non-billed accounts are also decentralized to individual cells. See cell budgets.
  • The sustainability role is responsible for ensuring the cell keeps the budgets it is responsible for in order.
  • Every 3 months, the sustainability manager posts an update on the forum about the current status of the cell's sustainability budget and ratio, as well as plans for the next months that ensures sustainable ratios. This is done in January (winter update), April (spring update), July (summer update), October (autumn update).

Cell manager - Prioritization of work#

  • The cell being the closest to the client needs, and knowing best the work to be performed, it's positioned the best to prioritize its own backlog accordingly. To do so, it must prioritize in the following order (decreasing priority):
    1. Client work to ensure that the client needs are met, and that they are happy with our work.
    2. Newcomer-friendly tasks should be ranked high enough in the list of tasks to allocate time for core team member review/mentoring. The earlier we can get newcomers up to speed, the better the workload will be for everyone.
    3. Internal projects flagged by the management as priorities.
    4. Additional work the cell finds useful to its or OpenCraft's function (or the Open edX community's), as long as the cell remains sustainable in proportion of billable hours. The cell defines and prioritizes the additional internal work, without a specific monthly budget cap (epic time budgets remain a good practice to follow, but the amount set doesn't require approval from management).

Note: It's important to keep in mind that to remain capable of leading initiatives at the company level, not just at the cell level, hierarchy can retain an important role in prioritization. Anyone in a cell can propose or prioritize a company-wide initiative, provided they get approval and buy-in from people it would affect. It would escalate to management if there's a conflict or disagreement about that, for example about how priorities relate to each other.

Client owner#

  • Client owners manage their relationship with their clients. All the epics and tasks from a given client are done within a single cell, the one the client owner belongs to, but see cross-cell collaboration for some nuances around task reviews.
  • Initial contact with prospects, estimations work. Note that a portion of Gabriel's time is assigned to each cell to do most of the initial contact and quote work with prospects.

Code reviewer#

As a code reviewer:

  • I will give prompt, thoughtful, thorough, constructive code reviews.
  • (When reviewing OpenCraft work): I will expect that the PR author has done everything outlined in the PR expectations part of the pull request process section - if not, I will ask them to fix that before I start the review.
  • (When reviewing non-OpenCraft work): If I get pinged to review a PR, I will respond to it within 24 hours. If it is for a ticket in the current sprint, I will ensure that I review it within 24 hours, or at least indicate to the author when they can expect a review.
  • I will aim to minimize review cycles (especially when reviewing non-OpenCraft PRs) by leaving as complete a review as possible. Ideally it should point authors to the exact changes needed for their PR to be accepted.
  • When reviewing work (or discussing it with the assignee before coding even begins), I will make sure that the assignee is not introducing code drift, i.e. that everything which could be upstreamed is (planned to be) upstreamed as part of the work.
  • I will always read through the code diff, considering things like:
    • Is the code easily understandable by humans?
    • Is the code of high quality?
    • Are all relevant coding standards followed?
    • Are all UX / a11y / i18n considerations addressed?
    • Is this introducing tech debt?
    • Is the new code well covered by tests?
  • I will always test the code manually, either locally or on a sandbox, unless there is no reasonable way to do so.
  • I will always check if any updates to the software's documentation are required, and either ask the author to update the docs, or ensure the relevant documentation team is pinged.
  • If there is any part of the code that I am not confident in reviewing, I will indicate that in my comments.
  • If there is any part of the code or PR that I particularly like, I will say so - we want to reinforce good practice as well as flagging issues.
  • I will set up the following template as a "Saved Reply" in GitHub and use it for the "stamp of approval" on all PRs I review:

- [ ] I tested this: (describe what you tested)
- [ ] I read through the code
- [ ] I checked for accessibility issues
- [ ] Includes documentation
- [ ] I made sure any change in configuration variables is reflected in the corresponding client's `configuration-secure` repository.

Here is a screenshot showing how to conveniently access this template when completing a code review using the dropdown button in the top right of the review panel (on the "Conversation" tab only):

Review template

  • If I'm the assigned reviewer for someone who is new to OpenCraft, I will reserve extra time to provide additional mentoring and support, and I will be especially responsive to questions. I will also provide feedback to the newcomer which will assist them during their trial period, and raise any major issues with their mentor.
  • If the assignee is clearly behind the schedule and doesn't respond at all to pings on the ticket or Mattermost within 48 hours, the code reviewer should determine whether this ticket is urgent. If it is, ask the firefighter for help to reduce the risk of spillover.


If I'm new to the team, I will:

  • Not commit to more than 3 story points of tasks during my first week, or 5 points my second week (so no more than a total of 8 points for the first sprint). It's really easy to over-commit at first, so keep your commitments manageable. You can always take on more work later in the sprint if you finish early.
  • Discuss general issues like time management and sprint planning with my mentor.
  • Tag my task reviewer(s) with task-specific questions on the ticket, or if I'm completely blocked, try using Mattermost to contact my reviewer, mentor, sprint firefighter, or other people working on the same project. Reviewers have time allocated to help during Onboarding, so it's ok to reach out!
  • Ask my Reviewer questions early in the sprint if there is missing information or if the requirements are unclear. Newcomer-friendly tasks have a minimum set of information that should be included in the task description, but if this isn't complete, ask the task creator and/or your reviewer to do it.
  • Provide lots of updates to my assigned reviewer/mentor for each task, so they can provide timely help.
  • Assign myself as reviewer for core team member tasks to learn about the various projects and systems that we support.
  • Log time spent "onboarding" (reading documentation, learning new systems, setting up devstacks) to my onboarding task. Time spent completing the work described in a task can be logged to the task itself, but please take care not to exceed the estimated time without approval from the owner of the parent epic. Some leeway will likely exist within the epic budget, but it's best to check to be sure.

Also, it is not compulsory to log X hours per week as stated in the contract -- our contracts give an estimated amount of time expected per week, but the actual work required for each sprint can vary due to many factors.

See the Process: Onboarding & Evaluation page for more advice on the onboarding, the evaluation process and criteria.

Onboarding epic ownership#

As a newcomers, I am the owner of my onboarding project, and will:

  • Keep time spent by me and my mentor on the onboarding epic to less than 80 hours for the initial 2 month trial period.

  • The remaining 100 hours of my onboarding budget must be preserved for the core team reviews, and any extended trial and/or onboarding time once accepted to the core team.

Expect around 15 hours to be spent by the core team on the screening and end-of-trial reviews during the initial 2 month trial period. For example, if the screening review requires 2 hours and the end of trial review requires 2 hours for each core team member and there are 6 of them, 14 hours (2 + 6*2) is required for those activities.

  • Treat my onboarding epic as a proper epic and manage it like an epic owner.

The epic and the budget also includes my onboarding ticket, my mentor's mentoring ticket and any other tickets created specifically for my onboarding like the ones for my screening review and the end-of-trial review.

In the onboarding epic task, there is a Summary Panel section on the right side, which shows various details about the budget, logged and available time etc. See the below screenshot for an example.

Summary Panel

Hover over the Time numbers below the progress bar to see an explanation.

  • I will create tasks under this epic for distinct pieces of work, set estimates and schedule them with the help of my mentor.

This will help us to track if a particular onboarding/learning task is taking too much time. Since the onboarding budget is limited, all the learning time I log on the onboarding epic will be related to the tasks that I work on.

  • Log my time with detailed descriptions about the work performed. For example:

    "Reviewed task and asked questions on ticket, requested repo access, started setting up devstack."

    Or as an example where logging "onboarding" time spent on another task on my onboarding ticket:

    "Onboarding for SE-123: read XBlock Tutorial and configured XBlocks in my devstack"


  • I will make sure to inform the team I'll mentor a newcomer before their first day.
  • I will familiarize myself with the current onboarding & evaluation process, so I can answer questions and help the newcomer through the process.
  • I will allocate sufficient time in my sprint, especially at the beginning, to assist and mentor the newcomer.
  • Before the newcomer's first day, I'll assign him/her a small, newcomer-friendly task for the first week as part of the screening process.
  • I will help the new member to find tasks that are suitable for someone starting up, especially in the first sprint and in the absence of Newcomer-friendly tasks.
  • During the trial period, I will help the newcomer select tasks to showcase their skills and learning abilities, with an eye to assigning tasks of increasing difficulty as the trial period progresses.
  • If newcomer-friendly tasks are not available in the current sprint or backlog, I will work with the epic owners to split out newcomer tasks from other work.
  • Alternatively, I can create one from the INCR project (preferable) or Byte-Sized Bugs upstream boards.
  • If no newcomer tasks can be created, I may assign a non-newcomer task. In these cases, I will consider factors such as (i) the circumstances of the task, (ii) the trust in the newcomer and (iii) the level of access needed for deployment. (Note: any additional access granted needs to be documented in the Access Doc)
  • I may assign very small discoveries, which do not result in a standalone epic. (See Newcomer-friendly tasks discussion)
  • If a task requires deployment, one option is that the reviewer does the deployment, but lets the newcomer know what they’re doing at each step, to help them learn for the future.
  • I will schedule an initial 121 meeting with the newcomer, and a recurring meeting every week. These can be spaced out later on if the meetings become less useful. Refer to the mentor checklists page for the topics to go through.
  • For the first 121 with the newcomer, I'll prepare a few interview questions to help in the screening process.
  • I will participate on the screening and developer review tasks, taking into account these evaluation criteria.
  • I will evaluate the newcomer's usage of the onboarding budget and raise any issues with Xavier. I'll also help the newcomer with creating, estimating and scheduling learning tasks in the onboarding epic, tracking them and with precise time tracking of his/her tasks if the work logs show too many approximate (i.e. rounded) values.
  • I will also explain to the newcomer, the guidelines to manage the onboarding epic and its budget like any other epic that we work on and point them to the relevant items in the newcomer section.


There are two firefighters per cell for each sprint and they are designated as Firefighter 1 and 2. Besides minor differences who leads which sprint meeting, these roles have exactly the same responsibilities.

As a sprint firefighter:

  • I will keep at least 15 hours of time for sprint firefighter duties as described below, and will proactively pursue those duties.
  • I will assign tasks for the rest of my sprint hours normally, but I will also ensure a few additional tasks are left either in the "stretch goals" or the following sprint during sprint planning, in case there isn't enough firefighting work to fill my hours. I will only pull these tasks into the current sprint if I am confident that I will have time to finish them in addition to the firefighting. These tasks should be assigned to me and have a reviewer, to be ready to pull in.
  • I will not record any time on the Sprint Firefighter task directly (always use a dedicated client story/bug/epic ticket).
  • I will be subscribed to the help and urgent mailing lists (be sure to filter them to a separate folder to look at them only when you need to - but also make sure that if any such emails also explicitly include you in the To/CC, then they will arrive in your inbox).
  • I will add myself to the pager rotation, at least during my work hours the weeks I am firefighter, plus any additional time I optionally want to help cover.
  • I will work on the following, listed by decreasing priority:
    • Handle emergencies from other team members (reported by team members directly to me or to the other sprint firefighter, with Braden arbitrating priorities)
    • Handle emergencies reported by clients from the cell--these will come in through the mailing list, which everyone receives. Firefighters triage and prioritize these issues, asking other cell members for advice as needed, and escalating to Braden/Xavier if everything else fails.
    • Handle critical bugs reported by QA
    • Deploying hotfixes/security fixes on client instances
    • Watch relevant forum threads (currently: Open edX devstack and Ocim sandbox issues and New periodic build instances (Ironwood and Juniper)) for reports about breakages of periodic builds and follow up by:
      • Determining if it is necessary to fix the breakage right away (e.g. because it's blocking some other work in the current sprint).
      • Creating a ticket in our internal JIRA and work on fixing the issue (or at least put in place a workaround), if it is in fact critical.
      • Report the breakage as a CRI issue otherwise (and ping Ned). See CRI-206 for an example (reported via SE-2587).
    • Provide reviews for tasks that are missing a reviewer.
    • Complete any personal spillover from the previous sprint
    • Work on client requests that can't wait until the next sprint, in particular in the first week of a sprint.
    • Help ensuring a clean sprint by helping other team members, in particular in the second week of a sprint.
    • Being available on the Open edX Slack (#general), answering questions and responding to emergencies reported there (log time on OC-1017)
    • Watch over sandboxes (ie keep an eye on the instance manager on a regular basis, few times a day, to ensure they build correctly, and debugging if needed)
    • Document incidents as they happen and post a summary to the ops review forum thread.

Dependency upgrades for security issues

GitHub sends the security alerts by scanning (by default, all public repositories are scanned; for private repositories, the admins have to explicitly permit the scanning) the repository dependencies, and based on the severity levels:

  • Critical and High level severity vulnerabilities: As there could be a risk of compromise or significant downtime for users, these vulnerabilities must be patched as soon as possible. The FFs should create the tasks for applying the patch and start working on them right away.
  • Medium level severity vulnerabilities: Report them and create tasks for patching them so that team is aware of them. These tasks can be scheduled for the next sprint and if any of the FFs have time, they can work ahead on these tasks.
  • Low level severity vulnerabilities: These vulnerabilities should be reported and tasks for patching them created with a lower priority. These tasks can be scheduled for a future sprint and prioritized by the epic owner as appropriate.
  • Undefined or unclear severity: Vulnerabilities or security fixes whose severity is unknown should be reported and discussed with team before we can take any action.
  • Dependabot PRs: Dependabot PRs for bumping the version of vulnerable dependencies or security fixes should be handled by the FFs based on the process mentioned above for various severity levels.

The severity level mentioned in a vulnerability report may not always match our own assessment (critical issues in a disabled feature may not be critical). Hence the processes around categorization of the security vulnerabilities and the assessment of their severity levels need to be expanded and improved over time.

  • Work on additional tasks from the backlog if additional time is available
  • Well before the sprint planning meeting at the start of a new sprint (every second Monday), I will remind the epic owners to get every ticket in the upcoming sprint "green" (has an assignee, reviewer, and points) by pinging people as needed, and I will help make sure all tickets are green before the start of the sprint planning meeting. If my timezone makes this challenging, I will coordinate with Firefighter 2 to finish up on Monday. The time spent getting a ticket to green will be logged on the ticket.
  • If I am Firefighter 1, I will lead the sprint planning meeting at the start of the sprint. If I am Firefighter 2, I will take notes during this meeting. For the mid-sprint meeting these roles are reversed, Firefighter 2 will lead the meeting and Firefighter 1 will take notes.
  • On Friday or Monday well before the mid-sprint meeting, I will do a sprint checkin on all "backlog" and "In progress" tickets (this means to post a comment on each ticket to ask how the assignee is doing and if they need help, and/or asking them to update the status of the ticket if it is not clear).

Community Relations#

General support for the community#

  • Rotating role, assigned to two cell members each sprint.
  • Keep 1h/sprint for providing support to the Open edX community via the following channels:
  • In general:
    • Monitor these channels for technical questions (e.g. about our XBlocks) and answer them. Generally favor helping one person more fully, over having many smaller contributions to more threads. It should give a feeling of what a support task with a 1h timebox gives. If it takes less than 1h to solve, you can help someone else with another problem with the remaining time. But otherwise spend that time on a single question/issue - one that you think you can contribute something useful to many, or that you're especially suited for.
    • If you run out of time, simply stop answering - someone else from the community can take over.
    • If you think it would be helpful to go beyond 1h in some cases, check a timebox extension with Xavier.
  • For Slack:
    • To make it easier to stay in the loop on what's happening there, consider changing your Slack settings for the #opencraft channel to "Notify on every message" while on rotation for this role.
    • If edX reports any urgent issues via the #opencraft channel, make sure that the firefighters are aware so these issues can be followed up on in a timely manner.
  • Time must be logged on a ticket dedicated to the role each week - see SE-1567 or BB-1692 for examples.

Official forum moderator#

  • As a moderator, edX is looking for help with:
    • Welcome new posters
    • Help people use the forums well, including:
      • Choose the right categories, and move/split/concatenate threads when needed.
      • Provide enough information about their situation - we can help them to refine their questions, asking them questions to make their requests more precise
      • Direct them to existing resources
    • Review posts flagged automatically by Discourse (for example, someone finishing a post too quickly after creating their account)
    • Set the right tone for discussions and be model Open edX citizens
  • This isn't about knowing the answers to questions - though it never hurts, and you can also provide support at the same time if it makes sense. But Ned defines it as "being involved with the forum in a way that makes it productive and useful for people."
  • The budget is 1h/week, in total between all the forum moderators

OSPR Liaison#

  • Keep 1h/week available to prioritize Open OSPRs for review by edX.
  • Maintain the prioritized list in a document shared with the OSCM team: OSPR priorities from OpenCraft
  • Send the top 3 on the list to the edX OSCM team (usually Natalia) once a week.
  • Log time on SE-1631.
  • Keeep 0.5hrs/week available to handle any PR sandbox requests by edX.
  • Someone (likely Ned or Natalia) will ping you on a PR targeting an older release branch (currently limited to Hawthorn and later).
  • Make sure the PR has a reviewer assigned. If not, request that one be assigned before continuing with the next step.
  • Create a new sandbox for the PR, and link the sandbox URLs to the PR.
  • To do this you can use the setup_pr_sandbox Ocim management command and provide it the URL of the edx-platform PR.
  • In case there isn't sufficient time, create a ticket and ping the firefighter to do the actual sandbox creation steps.
  • You can give SSH access to the reviewer by adding their GitHub ID to COMMON_USER_INFO in the PR sandbox's config:
  - github: true
    name: github_id
    type: admin
  • You can ask the PR author to add the standard setting section to the PR description with the above config. Otherwise you may need to add this back each time the sandbox is updated from the PR.

Ops reviewer#

  1. Receive and the pager alerts - check alerts/emails to ensure the alerts have been properly handled by the other recipients and nothing has slipped through.
  2. Be a backup on the pager (alerts sent first to the other(s) recipient(s), but escalated if not acknowledged within 30 minutes).
  3. If none of the other recipients are around and an issue is left unsolved, and it either is sent to urgent@, or was sent more than 12h ago to and needs to handled quickly, warn the sprint firefighter about it: create a ticket about the issue and assign it.
  4. Report breakages of periodic builds on the forum, via appropriate threads (currently: Open edX devstack and Ocim sandbox issues and New periodic build instances (Ironwood and Juniper)), and ping the firefighters for the current sprint for follow-up.

Discovery duty assignee#

As the week's discovery duty assignee:

  • I will make sure I am the right person to take on a specific discovery, if I feel I am not I will exchange work with other people from the sprint, to get someone else to do it, while I help them with their tasks. It is very important to be able to estimate a task well, we should make a conscious effort to assign the task to the person who can best handle it, if possible (if the person is available and willing to do the work).
  • I will keep at least 5 hours of my time to do discovery work on any small leads/client requests that come up during the week. I may still plan some tasks to pull into the sprint if no fires pop up. Before pulling tasks in the sprint, I will use these hours to help others finishing their sprint commitments.
  • If I do some discovery tasks, I will ping the sprint firefighters to review the result.
  • As always, if I do the discovery, I understand that I would be expected to complete the epic work later on if the client accepts it. (To avoid people being forced to commit to being the owner of big epics, any huge discovery stories should always be scheduled into future sprints rather than directly assigned to the discovery duty assignee.)
  • I will ensure an additional task is left either in the "stretch goals" or the following sprint during sprint planning, in case there isn't enough discovery duty work to fill my hours. I will only pull this task into the current sprint if I am confident that I will have time to finish it in addition to the discovery duty. This task should be assigned to me and have a reviewer, to be ready to pull in.

Specialist roles#

Specialists are people in the cell with more context and understanding about a certain topic, such as DevOps, Ocim, or Analytics/Insights. The goal of having this role is to allow cells to self-sustain and run its own projects without being blocked on context from people from another cell.

While specialists have priority on taking tasks related to its specialty, there's no restriction on other team members and taking tasks is encouraged to spread knowledge around the cell. Specialists should always be at least the 2nd reviewer so they are in a position to track the ticket and provide help.

Specialists are not only responsible for handling tasks within the cell, but also for coordinating with other specialists to discuss and schedule improvements when necessary.

The next sections cover specific tasks to be done by each type of specialist:

DevOps Specialist#

The specialist should always be involved in some way in the devops tasks of the cell, if not as the assignee, then as a reviewer or 2nd reviewer.

They also have priority to take DevOps tasks as assignee as much as they want, they don't have to take everything (neither should they try), but they should be able to use the context of tasks, especially client tasks, to implement useful work for the infrastructure as a whole.

Client owners are still responsible for handling DevOps tasks of their client's instance. The specialist should help, advise and review the tasks, while the epic/client owner creates the tasks and handles getting them assigned.

Responsibilities within its cell:

  1. Assign, take or review DevOps related tasks.
  2. Support team members on DevOps related tasks, provide information about deployment mechanisms, sidecar services, and any particular detail that might affect the outcome of the task.

Responsibilities shared between DevOps specialists across cells:

  1. Create tickets to maintain and update common server infrastructure and its related deployment mechanisms (MySQL, MongoDB, Consul, Vault, Prometheus, etc).
  2. Create discovery tickets and epics to improve infrastructure, taking into account the current epic schedule and throughput available for non billable DevOps work.

Technical Security Manager#

  • Responsible for the technical security of the work we do, our platform and OpenCraft as a whole.
  • Responsible for OpenCraft's Security policy.
  • Offboarding of team members who have left.

Epic owner#

As an epic owner:

  • I will make sure the epic has tasks with time estimates and a global budget. A discovery will likely be required for this - read the estimates section and follow the steps listed there.
  • I will ensure the "Timeline" section of the description and the epic due date are kept up to date. The "Timeline" section should include a list of milestones and deadlines for each (these milestones could be as simple as "check in with client"). The due date should be set to the closest deadline, and must be revised once each deadline is past/updated.
  • I will keep clients updated on the status of tickets (for clients with long term monthly contracts, this means updating their JIRA or Trello boards as work progresses).
  • I will attend meetings with the client about the project as required, e.g. weekly scrum meetings. (We try to limit these sort of meetings to once a week per client at the most.)
  • I will create tickets on OpenCraft's JIRA board for upcoming work, and place them into the appropriate upcoming sprints so that we can get the work done well ahead of applicable deadlines.
    • If a ticket without point estimates is scheduled for the upcoming sprint, and an estimation session is open, I will make sure to add it to that session, so everyone on the team can help estimate it.
  • I will aim to split out any suitable newcomer-friendly work into separate tasks, flag them as Newcomer-friendly tasks in the Backlog, and ensure they contain the required information. To guard the epic budget, "onboarding" time can be logged against the newcomer's epic.
  • If applicable, I will create corresponding tickets on clients' JIRA or Trello boards and link to those tickets from the description of internal tickets as necessary. (In many cases, a single internal ticket from OpenCraft's JIRA will correspond to a single ticket from a client's JIRA or Trello board. But there are also cases where a single internal ticket might map to multiple external tickets or vice versa.)
  • I will set the "Original Estimate" of each ticket to the number of hours listed in the discovery document. If a ticket covers a larger story from the discovery document only partially, I will set its "Original Estimate" to an appropriate fraction of the number of hours listed in the discovery document. For any ticket that does not directly correspond to a story from the discovery document, I will set the "Original Estimate" based on the amount of time that I think will be required to complete the work, taking into account the number of hours that remain in the epic's budget at that point in time.
  • On Thursday in the second week of a sprint, I will post an epic update that answers these questions:
h5. Epic Update (Sprint ???)

h6. What's the status of the hours budget for this epic?

We've used __ hours of the __ hour budget.

h6. What's left to do? (Is it on track to be done before the deadline and within the hours budget?)


h6. Are the deadlines in the "Timeline" section of this epic's description correct?

(/) or (x)

h6. Are all upcoming stories that we should work on in the next week or two created and in the correct upcoming sprint?

(/) or (x)

h6. Is the status of each story in the epic correct?

(/) or (x)

h6. Other Notes


Other Notes is for any concerns, questions, upcoming vacations, or anything else that you think is worth mentioning because it could affect planning of the next sprint or two.

  • On Thursday or Friday in the final week of each sprint, I will assign upcoming stories from this epic that need to be done next sprint to myself - or I will ask others from my cell to be either the assignee or the reviewer. (If the epic owner is away, the epic reviewer is responsible for this.)

Client owner#

For each client, we have a single person who is designated as the owner for that client, called the "client owner." The client owner is responsible for handling most communications with that client. The client owner is the person who is assigned to the epic or task with the name "\ [General] Support", such as OC-2340 or OC-2910. If nobody is assigned to such an epic, or it doesn't exist, the epic owner of the current year's Small Projects & Customizations epic is the default client owner.

In general, regardless of which email address the client sent their question to, the client owner should be the one who replies, though they may hand off any conversation to anyone on the team as needed (such as their cell's firefighter). If someone else received the email, please "Reply All" to pass the message on to the client owner for them to reply (add the client owner to the "To" field, and add "OpenCraft" to the CC field, and say something like "Passing this on to (client owner name)").

There are a few exceptions:

  • If it's an urgent problem, whoever sees it first should reply and CC the firefighters,, and the client owner. One of the firefighters should respond.
  • If it's a general issue/question:
    • If the email is related to a specific project/epic with a different epic owner, that epic owner can reply directly (make sure the client owner and are CC'd).
    • If the client owner is away, their designated backup person (usually the reviewer on the "Support" epic/ticket for that client) should take over this role. (Or the firefighters if no backup person was planned.)

We have a few clients with monthly development budgets. For those clients, the assigned client owner is also responsible for reviewing the monthly budget before and during each sprint planning meeting, to ensure that we are not going too far above or below the budget. They should refer to the Sprints which can monitor and project the status of each budget as of the end of the upcoming sprint.

When we get a new client, or when the client owner changes (or we start working with a new person coordinating things on the client's side), we send them a welcome email explaining who the client owner is and how to contact us.

Product specialist#

As a product specialist:

  • I will ensure clear and prompt communication with prospects, leads and clients.
  • I will regularly look out for business opportunities. This includes consulting public tendering platforms, sending cold emails to prospects, and attending events where you could meet prospects. Strategy and time spent on these tasks should be discussed with the cell sponsoring the prospect work, and possibly Xavier for key prospects.
  • I will send the official OpenCraft offering presentation materials to prospects.
  • When meeting, I will describe OpenCraft’s offering, history and philosophy to prospects.
  • I will quickly gather prospects’ needs and potential time budget through discussion. This enables us to know if and how we can help them.
  • I will keep an open mind. Buyers might have needs that we can’t fulfill yet - this might be an opportunity and should be discussed with the cell, and possibly Xavier for key prospects.
  • When a prospect becomes a lead, I will follow-up by providing information, drafting quotes, creating epics and issues and documenting in Jira so the team can follow what’s happening. I will include the template in the epic description ("Context, Scope, Budget, Timeline") and set the time budget on the epic ticket.
  • Handle the onboarding of confirmed clients: tell them how to work with us, such as when/how to use the email address, when/how to use, who their main contact(s) should be, how billing works, etc.
  • When managing a project, I will handle communication and coordination between the client and the cell working on the project, and assign discovery tasks to developers until the client approves the quote.
  • Generally, when writing to the clients, I will always make sure there is either contact@ or help@ CC'd:
    • anything related to accounting/billing/quote should be on contact@
    • and the rest on help@
  • I will manage and monitor self-hosting registrations and addressing user issues in Ocim
  • Handling billing and accounting for self-hosted users
  • Handling contact@ email
  • OpenCraft lead management flowchart
  • OpenCraft client onboarding flowchart

Learning designer#

  • Providing clients with pedagogical advice and strategy
  • Providing clients with custom learning experience designs


  • Technical Arbiter
    • Final arbiter for technical decisions
  • Technical Excellence
    • Ensuring technical & product quality of everything we do
    • Ensuring software we produce is as open as possible and everything that can be contributed upstream is
    • Championing and improving processes (like our code review process) to improve the quality and efficiency of our software development
    • Reviewing code (in addition to the regular code review process), mentoring developers
    • Being aware of the "big picture" to help developers avoid duplicating each other's work
    • Being available on call 24/7 to respond to emergencies, if regular firefighters are not available
  • Epic/Client Owner Coach
    • Supporting epic and client owners in client meetings, discussions, and in turning requirements into actionable plans/epics
    • Reviewing all discovery stories and estimates
  • Security Manager
    • Ultimately responsible for security of our software, both proactively improving it and responding to incidents effectively

The CTO isn't part of a cell, but can review and participate in any cell's work as needed.

Sprint Manager#

Tasks shared between sprint managers from all cells:

  • Close the current sprint and start new one after sprint meeting.

Tasks to be done individually by each sprint manager:

  • Create refinement session on Wednesday before next sprint, and close it before the sprint meeting, allowing time for the assignees to take them and make proper planning.
  • When creating the session, add all the unestimated tickets for next sprint and on stretch goals to it.
  • Add epic owners as Scrum Masters, so that they can add it the tickets they create or pull in after the estimation session was created.
  • Ping everyone that hasn't participated on the estimation session to estimate before closing the session.
  • Create and assign Firefighter and Discovery Duty tickets for upcoming sprint.
  • Review of sprint deliveries: Move tasks from "Deployed & Delivered" to "Done", after double-checking that all the criteria for calling it "Done" have been met.
  • Review of sprint planning & tickets in "waiting for 3rd party" and "long reviews".
  • Update the sprint commitments spreadsheet at end of sprint.


  • Business development:
    • Additional relationship with key clients and prospects
    • Support role for the epic owner, in difficult situations or contract negotiations
    • Reviewer of quotes
    • Initiator for opportunities that are strategically important for OpenCraft, either because of size or side effects
  • Financials:
    • Accounting
    • Invoicing (clients & team)
    • Forecasting
  • Admin:
    • Holidays reviews
    • Partial offboarding (main work done by CTO)
  • Legal:
    • Relationship with lawyers
    • Contracts drafting and signature (team and clients)
    • Ownership of legal projects
  • Management:
    • Management of non-technical matters for all cells
    • Management of CTO
    • 121 meetings: continuous rotation of all employees, at 2-3 meetings/week
    • Handle reports of performance issues & process to address it, including firing if needed
  • Hiring:
    • Sourcing of initial candidates, upon demand by individual cells
    • Contract negotiation with accepted candidates
    • Granting access to new hires & core developers
  • Sprint management:
    • Occasional second review of sprint deliverables, sprint planning & epics management on any cell
  • Strategy: evolution of the company
  • Ocim: product management and some code
  • Spreading the word (mailing lists, conferences,