REKKY first. Mission complete.
On-call and incident process
It's done.
The rotation, the page, the note, and the review the next week. Built so the people you already have can carry the night. Not a retainer.
Four pieces, in this order
On-call and incident process work has four pieces. A rotation that names a person. A page that person can act on. A note written while the facts are fresh. A review that changes the next week, or explicitly declines to. Companies often have a chat channel and a hope. The engagement writes the four pieces in the tools you already use.
The order matters. A beautiful review of an alert nobody could act on does not fix the night. A rotation with no backup does not survive a vacation. We start from the last real incident or the last useless page, not from a blank policy.
This page used to be framed as readiness beside a wider operations offer. It is now the incident process for this DevOps and SRE practice. Response retainers are still not what is sold.
The rotation you actually have
We read the schedule as it was kept, not as it was announced. Who was primary, who was backup, who was both because the bench is one person, and what happened when that person was asleep or out. A schedule that exists only in a spreadsheet nobody opens is a finding.
Load is part of the picture. How many pages in a recent stretch, which of them were acted on, and which were silenced. A rotation that trains people to ignore the phone is not a rotation. We will say that plainly.
Handing the phone to the consultant is not the design. The people who own the service stay on it. We need them in the room while the rotation is rewritten. We run the engagement end to end and work around your normal operations, but we'll need access to the right areas and the right people while we do.
The page and the note
A useful page names the service, what looks wrong, and a first place to look. A page that only says a threshold was crossed, with no service and no recent change, is noise. Noise is removed or rewritten with the person who receives it. We do not add alerts to look thorough.
The note answers a small set of questions. What did users hit, what changed, what was tried, what remains. It lives where the next on-call will look. A document store nobody opens is not a record. We do not sell a form. The questions are fitted to your service.
- Primary and backup, as last month was actually staffed.
- Pages from a recent stretch, split into action and noise.
- One note from a real incident, rewritten only where the next person would get stuck.
- A review that either changes a page, a run step, or the rotation, or says why it will not.
What we will not do
We will not take the pages as a retainer. We will not staff your nights. If the message you want to send is that you are down and need someone on the keyboard tonight, send it plainly. We will tell you if that is not this engagement.
We will not write a policy that your team has not tried. A step that only works if everyone is calm and fully staffed does not go in the note. The test is a quiet week: can the backup follow the note without a phone call to the consultant.
SRE consulting covers the service itself, not only the night around it. Use that page when the question is whether the service can stay up. Use this page when the question is whether the night is survivable. The Recon Assessment, starting at $15,000 plus expenses, can include both when you want them in one report.
Questions on this work
Will you take our pages?
Not as a retainer. The work puts a rotation, a useful page, a note, and a review in place so the people you already have can carry the night. If you need someone on the keyboard tonight, say that. We will tell you if that is a different firm.
What is a useful page?
A page that a person can act on: which service, what changed, and who can roll it back. An alert that only says the graph is red is a candidate to fix or remove.
Does the incident note have a template you sell?
No. The note has to match how your team already talks about the service. We write the smallest record the next person can use, and you keep it.
Contact
Describe the platform you have.
Tell us who was paged last, and whether that person could act on what the page said. We run the engagement end to end and work around your normal operations, but we'll need access to the right areas and the right people while we do.
Recon Assessment, starting at $15,000 plus expenses. Four days on site with your team, then a written report within about two weeks.