Incident.io take-home exercise

Take home exercise at Incident.io.

On-Call scheduling take-home task

We expect this task to take you about an hour to complete.

Please feel free to use AI coding tools such as Cursor or Claude to complete the challenge, just as you would in a real-world scenario. However, the expectation is that you will fully understand the code you submit, and could comfortably explain it to a colleague or extend it.

Additionally, you can use any libraries that you choose.

Background

At incident.io, we believe everyone should be on-call by default.

It's also important that you can get 'cover' for parts of your on-call shift, for example if you want to go for a swim or you've had a bad night and need to catch up on sleep.

This is a script we use to figure out who was actually on-call when, based on a list of schedule entries (the original plan) and overrides (when someone else temporarily took the pager).

The Challenge

We'd like you to 'fill in the blanks' in the script, so that when someone runs the script it will load the data from input.json and then:

1. Check that the schedule entries do not overlap with each other
2. Check that there are no 'gaps' in the schedule (i.e. no time when no one was scheduled to be on-call).
3. Check that the override entries do not overlap with each other
4. Flatten the schedule entries and overrides into a single list of 'entries', that represents who was actually on-call at any given time.

Output

The script should either:

1. Exit with an error if there are any gaps in the schedule, or overlapping overrides
2. Write an output.json file in the data directory that contains the flattened list of entries.

Assumptions

You can assume that:

• The input data structure is valid (JSON is well-formed, dates are in ISO 8601 format, etc.).
• There will be no more than 100 schedule entries or override entries.
• All dates are in UTC.
• Overrides will not extend beyond the start or end of the schedule.

For the purposes of this task, you do not need to consider performance optimisations: we're much more interested in the structure, correctness and clarity of your code.

Understanding the on-call schedule

An example schedule export can be found in input.json.

There are two lists of entries in the data:

• schedule_entries: This is a list of periods when a user is scheduled to be on-call. There should be no gaps in this list, meaning that one shift starts as the one before it ends.
• override_entries: This is a list of periods when a user temporarily takes over the pager from another user. These can overlap with the schedule entries, and should be compensated instead of the original user.

The input:

{
  "schedule_entries": [
    {
      "user_id": "1",
      "start_at": "2024-11-02T16:00:00Z",
      "end_at": "2024-11-09T16:00:00Z"
    },
    ...
  ],
  "override_entries": [
    {
      "user_id": "3",
      "start_at": "2024-11-09T15:00:00Z",
      "end_at": "2024-11-09T20:00:00Z"
    },
    ...
  ]
}