Octri

monitoring slos

An SLO is a target for one transaction: a p95 latency ceiling, a maximum error rate, or both. This lists every transaction with a target and judges it on the window: its p95 against the ceiling, its error rate against the maximum, and how much of the error budget is spent. See service level objectives.

A target only shows once its transaction has traffic in the window and is among the 200 busiest.

bash
octri monitoring slos [list | set <transaction> | delete <transaction>] [--fail]

Flags

--range <window>stringdefault: 24h

The time window: 1h, 6h, 24h, 7d, 30d or 90d.

--failboolean

Exits with an error while any target is breached, so a deploy job can check it after release.

octri monitoring slos set <transaction>

Adds a target, or replaces the one there. Give --p95, --error-rate, or both.

Arguments

<transaction>stringrequired

The transaction as monitoring names it, in quotes, such as "GET /users/:id".

Flags

--p95 <ms>number

The p95 latency ceiling, in milliseconds.

--error-rate <percent>number

The maximum error rate as a percent, so 1 means 1%.

octri monitoring slos delete <transaction>

Removes the target.

Arguments

<transaction>stringrequired

The transaction, in quotes.

Examples

Set a target

bash
octri monitoring slos set "GET /customers" --p95 300 --error-rate 1
Output
✔ Setting the SLO for GET /customers
✔ SLO set for GET /customers
  octri monitoring slos: how it stands on the current window

Check targets after a deploy

bash
octri monitoring slos --range 1h --fail