Content & Marketing
Technical Writer
Technical Writer turns rough feature notes, change summaries, and implementation context into documentation people can actually use. It supports architecture notes, API references, runbooks, user guides, and changelogs with a level of structure that saves teams real time. Engineering teams, developer experience groups, product managers, and agencies benefit when documentation quality matters but writing time is limited. The skill helps keep knowledge flowing across internal and external audiences without requiring every expert to become a documentation specialist. Its production-grade quality comes from breadth and discipline. It can generate multiple documentation forms from the same underlying context, which makes it valuable as an operational writing tool rather than a one-purpose prompt.
One-Time Purchase
$19.99
Deployment Runbook: Order Processing Service
Type: Runbook Audience: On-call SRE and backend engineers Owner: Platform Team Service criticality: Tier 1 — revenue-impacting
What this runbook covers
Standard deploy, canary deploy, and rollback for the order-processor service. The service sits on the synchronous checkout path — every minute of downtime correlates with measurable revenue loss, so the pre-flight gate is intentionally strict. If any pre-flight check fails, do not proceed; page the on-call platform lead instead.
Pre-flight Checks
| # | Check | Severity if missed | How to verify |
|---|---|---|---|
| 1 | AWS CLI v2.15+ on prod-deploy profile | Blocks deploy | aws sts get-caller-identity --profile prod-deploy |
| 2 | kubectl context = prod-us-east-1 | Wrong-cluster deploy risk | kubectl config current-context |
| 3 | Membership in #deploy-notifications | Communication only | Open Slack, confirm channel |
| 4 | ArgoCD session active | Blocks deploy | https://argocd.internal.example.com (must show your username) |
| 5 | Error rate < 0.1% for last 15 min | Do not deploy into a degraded service | Prometheus query below |
| 6 | No active SEV1/SEV2 incidents | Deploys frozen during incident | PagerDuty incident dashboard |
# Verify current running version
kubectl get deployment order-processor -n orders \
-o jsonpath='{.spec.template.spec.containers[0].image}'
# Expected: registry.example.com/order-processor:v2.14.3
# Current error rate (should be < 0.001)
curl -s 'https://prometheus.internal/api/v1/query?query=rate(http_errors_total{service="order-processor"}[5m])' \
| jq '.data.result[0].value[1]'
Gate — do not bypass
If error rate is above 0.005 or any pre-flight check is red, abort the deploy and page @platform-oncall. We have shipped into a degraded service before; every time, it converted a small fire into a much larger one.
Standard Deployment
1. Trigger the deploy
argocd app sync order-processor --revision v2.15.0 --prune
2. Verify the rollout
kubectl rollout status deployment/order-processor -n orders --timeout=300s
3. Watch the post-deploy window
| Window | Watch | Healthy band |
|---|---|---|
| 0–5 min | Error rate, p99 latency, restart count | err < 0.001 · p99 < 400ms · 0 restarts |
| 5–30 min | Order-completion rate vs 24h baseline | Within ±2% |
| 30–60 min | Downstream-queue depth (payment-gateway-q) | No sustained growth |
Canary Deployment (optional, recommended for risky changes)
Canary windows
Promote between stages only if all three watchers in the previous window are green. Hold and assess before promoting if any single watcher is amber.
Rollback
# Roll back to the last known good revision
argocd app rollback order-processor
# Confirm the rollback landed
kubectl get deployment order-processor -n orders \
-o jsonpath='{.spec.template.spec.containers[0].image}'
Rollback is always the first response to a post-deploy regression. Diagnose after the customer impact is contained.
Known Failure Modes
Failure mode 1 — DB migration drift
Pods stuck in CrashLoopBackOff immediately post-deploy almost always indicate an unmigrated schema change. Run kubectl logs -n orders -l app=order-processor --tail=50 and look for migration errors. Do not retry the deploy — roll back, then resolve the migration drift before redeploying.
Failure mode 2 — unindexed-query 5xx spike
5xx spike that begins exactly at deploy time and tracks closely with traffic volume is almost always a new code path hitting an unindexed query. Roll back immediately. The fix is an index, not a query rewrite; add the index in a separate change and redeploy.
Failure mode 3 — payment-gateway queue backup
The payment-gateway-q depth growing for 10+ minutes post-deploy without an error-rate signal usually means a downstream timeout was lengthened in this release. Confirm with the payments team before rolling back — rollback in this case can make the backlog worse.
Failure mode 4 — slow rollout (image pull)
If kubectl rollout status hits the 300-second timeout, the most likely cause is image-pull throttling from the registry. Check kubectl describe pod for ImagePullBackOff; the fix is usually a registry-side rate limit, not the deploy itself.
Troubleshooting Reference
| Symptom | Likely cause | Resolution |
|---|---|---|
Pods stuck in CrashLoopBackOff | DB migration drift | Roll back, then resolve migration |
| 5xx spike at deploy time | New unindexed query | Roll back; add index in separate PR |
ImagePullBackOff | Registry throttling | Confirm with platform; usually clears in 5 min |
| Rollout stuck at "Progressing" past 5 min | Readiness probe failing | kubectl describe pod and read probe logs |
| p99 latency 2× baseline post-deploy | Cold cache | Hold and re-measure at 15 min; cache warms quickly |
Post-deploy Checklist
- Image SHA matches release manifest
- Error rate and p99 inside healthy band at 30 min
- Order-completion rate inside ±2% of 24h baseline
- Deploy noted in
#deploy-notificationswith the release tag - Linked Jira ticket moved to "Released"
This sample illustrates the skill's output format. Service names, registries, and dashboards are illustrative. ClearPoint Nexus is not affiliated with the companies named.
View full sample →
All sales final. No refunds on digital products.
Includes support for Claude Code, Codex, OpenClaw, and Google Antigravity in the same license.
Also in Content Publishing
Bundle price: $55. Compare this skill with the full workflow bundle or Pro access.
Best for
Engineering teams without a dedicated docs writer producing release notes, internal runbooks, and onboarding guides as a side-of-desk task. Especially valuable for DevRel groups translating internal context into external docs at the pace of product change, and for platform teams who need consistent voice across dozens of service READMEs.
Not ideal for
Customer-facing API reference docs that must be generated from the source-of-truth OpenAPI spec — for that, use the API Documentation Generator instead. Also a poor fit for tutorial content where the value comes from a specific author’s walkthrough style; the output is workmanlike, not personality-driven.
Included in this purchase
- Claude Code, Codex, OpenClaw, and Google Antigravity skill files.
- Setup guidance for the right adapter in your workspace.
- One-time license for the purchased skill version.
Setup
Plan for a short copy-and-configure setup in your preferred agent workspace. No custom integration is required for the skill file itself.
Related Skills
$19.99
One-time license
$19.99
One-time license
$19.99
One-time license
Future Updates
This purchase includes the current version of the skill. If you want future adapter updates — meaning compatibility and packaging updates as supported platforms evolve — plus new catalog additions included automatically, upgrade to Pro.