agentleFS
stevancrisstevancris/sre-ai-agent/skills/slo-monitoring/SKILL.md

slo-monitoring

Skill0 starsChanged 5 months agoMITView on GitHub

Define, review, and alert on Service Level Objectives and error budgets. Use when setting up SLOs, reviewing SLA compliance, calculating error budget burn rate, debugging why an error budget is depleted, or planning reliability investments. Trigger keywords: SLO, SLA, error budget, burn rate, reliability target, availability, latency percentile, p99, p95, p50, uptime, four nines, five nines, nine nines, service level objective, service level agreement, error budget exhausted, reliability investment.

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 6 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Before loading context files, check if context/CONTEXT.md exists in the current directory.

    line 17
  2. Distinctive:

    Use the answers inline for this session.

    line 29
  3. Very specific:

    Read context/CONTEXT.md and the company's SLO configuration from context/CONTEXT.md (primary_slo_target and error_budget_policy fields).

    line 38
  4. Distinctive:

    Look at historical performance over the last 90 days.

    line 65
  5. Distinctive:

    Ask the user for (or calculate from context):

    line 96
  6. Distinctive:

    When error budget is burning faster than expected:

    line 124
  7. Distinctive:

    Ask: what changed around that time?

    line 127
  8. Distinctive:

    Never set an SLO tighter than what you can currently measure accurately.

    line 138

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

SLO Monitoring Skill

Setup Check

Before loading context files, check if context/CONTEXT.md exists in the current directory.

If context/CONTEXT.md exists — read it and proceed normally.

If context/CONTEXT.md does not exist — this skill was installed standalone (e.g. via npx skills add). Ask the user these questions before proceeding:

  1. Role — junior-sre / senior-sre / sre-manager (shapes output depth and tone)
  2. Cloud provider — aws / gcp / azure / on-prem / hybrid
  3. Observability stack — e.g. Datadog, Prometheus+Grafana, New Relic
  4. Company name and primary services affected (if relevant to this task)

Use the answers inline for this session. For persistent setup across all skills, suggest:

pipx install sre-agent
sre-agent init

Instructions

Step 1: Load Context

Read context/CONTEXT.md and the company's SLO configuration from context/CONTEXT.md (primary_slo_target and error_budget_policy fields).

Step 2: Determine the Task

Classify the user's request into one of these modes:

  • Define SLO — no SLO exists yet for a service
  • Review SLO — existing SLO, check current status
  • Diagnose burn — error budget is depleting faster than expected
  • Policy review — is the error budget policy appropriate?

Mode: Define SLO

2a. Choose the SLI (Service Level Indicator)

For each service, identify the right SLI type:

Service typeRecommended SLI
Request-serving (API, web)Availability (successful requests / total requests)
Request-serving (latency-sensitive)Latency (% of requests under threshold)
Data pipelineFreshness (age of latest processed record)
StorageDurability (% of writes successfully persisted)
Batch jobCompletion rate (successful runs / total runs)

2b. Set the SLO Target

Use this framework:

  • Look at historical performance over the last 90 days.
  • Set the target slightly below historical performance (leave room for improvement).
  • Align with customer expectations (explicit SLA or implicit from product tier).

Common starting points:

  • Consumer-facing API: 99.9% (43.8 min/month downtime allowance)
  • Internal API: 99.5% (3.65 hours/month)
  • Batch pipeline: 99.0% (7.3 hours/month)

2c. Define Burn Rate Alerts

Set multi-window, multi-burn-rate alerts using these thresholds:

AlertBurn rateShort windowLong windowError budget consumed
P1 (page)14.4x1 hour5 minutes2% in 1 hour
P2 (page)6x6 hours30 minutes5% in 6 hours
P3 (ticket)3x3 days6 hours10% in 3 days
P4 (notify)1x30 days—budget on track to exhaust

2d. Define Error Budget Policy

- If error budget remaining > 50%: feature work proceeds normally
- If error budget remaining 10–50%: add reliability work to sprint
- If error budget remaining < 10%: freeze non-critical feature work
- If error budget exhausted: full reliability mode until budget replenishes

Mode: Review SLO

Ask the user for (or calculate from context):

  • Current SLO target (%)
  • Measurement window (rolling 30-day is standard)
  • Number of bad minutes or bad requests in the window

Calculate:

Error Budget = (1 - SLO_target) × window_duration
Error Budget Consumed = (1 - current_availability) × window_duration
Error Budget Remaining = Error Budget − Error Budget Consumed
Error Budget Remaining % = (Error Budget Remaining / Error Budget) × 100
Days Until Exhaustion = (Error Budget Remaining / current_burn_rate)

Output a status card:

SLO Status: <SERVICE NAME>
Target:     99.9%    (43.8 min/month error budget)
Current:    99.85%   (65.7 min consumed)
Remaining:  0% — BUDGET EXHAUSTED
Burn rate:  1.5x normal
Action:     Error budget policy: reliability mode

Mode: Diagnose Burn

When error budget is burning faster than expected:

  1. Identify the time range when burn accelerated.
  2. Ask: what changed around that time? (deploys, config changes, traffic spikes)
  3. Identify the specific failure mode: high error rate, high latency, or both?
  4. Calculate which service or endpoint is contributing most to the burn.
  5. Recommend: fix, freeze, or adjust target?

Guidelines

  • Tight SLOs (99.99%+) require significant engineering investment. Question whether the business actually needs that level and what the cost is.
  • Never set an SLO tighter than what you can currently measure accurately.
  • Error budgets are not punishments; they are a shared language for trading features vs. reliability.
  • Persona (sre-manager): always include a "business impact" section — what does this SLO mean for customer retention, SLA penalties, or brand risk?

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.