REKKY first. Mission complete.
SRE consulting
It's done.
Consulting for the reliability of services you already run, and for the on-call that has to carry them. Not a new department, and not a retainer.
What SRE consulting means here
SRE consulting on this site is a look at production as your on-call already experiences it. Which service customers notice, what failure looks like on a Tuesday, who gets the page, and whether the note from the last failure would help the next person. The vocabulary can stay plain. The point is the week, not a new title for the team.
If you already have a target for availability or latency, we use yours. We do not arrive with a number and ask the service to match it. If you have no target, the work writes what the service has to do in the words operators use, and what will be treated as a failure. That sentence is more useful than a dashboard nobody believes.
The practice is Coalesce Ops, inside REKKY Consulting LLC. Fractional DevOps consulting is the path to production. This page starts after the service is already there.
Reliability of what is already up
The review picks one service that matters and walks a normal hour and a bad hour. What the user hits, what it depends on, what changed last, and how someone knows it is sick before the customer says so. Dependencies that are invisible until they fail get named. A dependency with no owner is a finding, not a footnote.
We look at the signals you already collect before anyone proposes a new one. A graph that did not help during the last incident is not a reason to add three more. A missing signal is in scope when the last incident could not be seen with the tools you have.
Capacity, deploys, and configuration show up only as they touch that service. A general inventory of servers is not the engagement. Platform operations covers the path a change takes. This page asks whether the running service can be trusted on the week you have.
The on-call beside the service
Reliability work that ignores the rotation is a document. We read the rotation you have: who is primary, who is backup, what happens when the primary is the only person who understands the service, and which pages fired in a recent stretch. A page that cannot be acted on is a candidate to fix or remove.
The incident note is part of the same picture. Where it lives, whether the next on-call can find it, and whether it says what changed and what was rolled back. We do not sell a template. The note has to match the service.
- One production service, walked on a normal hour and a bad hour.
- The signal that did or did not show the last failure.
- The rotation as it is staffed, not as the chart wishes.
- The note the next person would actually read.
Joining the rotation is not the default. If the scope includes a week beside your primary, it says so, and someone who already carries the service is in that week. A rotation that only works while the consultant is listed is not a result.
Limits of this engagement
SRE consulting here is not a 24-hour response retainer. It does not answer the phone at 2 a.m. in place of your team. If you are in an active outage and need hands on a keyboard tonight, say that in the first text. We will tell you if that is a firm we are not.
It is not a promise of a reliability percentage and not a tool purchase. Changes to paging or to the service are made with the people who own them. 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.
A fractional period that is mostly release engineering belongs on the fractional DevOps page. A bill that does not match the service belongs on the cloud cost page. The Recon Assessment can include more than one of those when the scope says so. The Recon Assessment, starting at $15,000 plus expenses, is the published four-day form.
Questions on this work
Is SRE consulting a 24-hour response retainer?
No. SRE consulting here builds a reliability practice your team can run: what pages a person, what gets written down, and what is reviewed the next week. It does not answer the phone at 2 a.m. as a retainer.
Do you invent an error-budget number?
No. If you already have a target, we use it. If you do not, the work names what the service has to do on a normal week and what failure looks like, in the words your on-call already uses.
Will the consultant join the rotation?
Only if the written scope says so, and only beside someone who already carries the service. A rotation that depends on the consultant after the engagement ends is not finished.
Contact
Describe the platform you have.
Name the service customers notice, and name who was paged the last time it failed. 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.