Saint Pierre and Miquelon

Just off the southern coast of Labrador, there’s a cluster of small islands named Saint Pierre and Miquelon.

Visually, there’s little to distinguish these rocky, scrub-covered islands from the other rocky, scrub-covered islands around Labrador. But beneath their unremarkable exterior, these islands are a historical oddity: they are technically part of France.

Due to one of those quirks of history, they remained with France even as the rest of Canada was ceded to Britain. Some highlights of the Wikipedia summary of their 300-year history:

• The British invaded and occupied the islands during the American Revolution, the French Revolution, and the 100 Days War, deporting the whole population of the islands each time
• The islands were responsible for smuggling almost 2 million gallons of whisky into the US in a 12-month period during Prohibition
• The Free French invaded and conquered the islands during WWII after the local government sided with the Vichy regime

But in my day-to-day life, the most important thing about Saint Pierre and Miquelon is its flag:

Or rather, the emoji version of its flag: 🇵🇲

The Slack shortcode for this flag emoji is :flag-pm:, which is a very useful tool during incident response.

As a brief aside, Sigma’s incident response automation will automatically set up a dedicated Slack channel, Zoom room, and postmortem document when a new SEV1 or SEV0 incident is reported via Jira. All discussion for the incident happens in the dedicated Slack channel or Zoom room, which keeps incident-specific discussion out of our other channels and provides an easy record of our incident response for when we conduct the postmortem. Once the incident is resolved, we write up a postmortem (PM) document and conduct a blameless postmortem review meeting to detemine what we can do better in the future.

In the heat of the incident, things will often come up in the Slack discussion that we’ll want to cover during the postmortem. Some recent examples:

• a teammate didn’t have access to an internal organization in the cloud region that was experiencing the outage
• our trace logs were missing a piece of data that would have helped us falsify a hypothesis about the incident cause
• while investigating a particular database query that was throwing an error, we noticed that the overal CPU usage of our database had been elevated for the past two days

Rather than adding a new comment to the Slack channel to say “I think we should discuss this in the postmortem”, we have a convention of reacting with the :flag-pm: emoji (as in, “I want to flag this for the PM”). This has the following advantages:

• it’s faster than typing out a new comment, particularly when the conversation is moving quickly
• it provides less noise in the channel than adding separate comments
• there’s a single, unambiguous way to search for all such flagged comments (from the Slack command bar, search in:#incident-channel has::flag-pm:)

It’s a small thing, but we’ve found it provides a quality of life improvement during stressful moments when time is at a premium.