Usage-priced log platforms meter every gigabyte — and the surprise-bill stories are public. The free parts (Vector, Loki, Grafana) are excellent but ship unassembled: Loki has no authentication layer, no retention policy, no alerting. This kit is the assembly: one zip, one command, and the whole stack runs on your own VPS — auth, retention and alerting included from the first boot. $59 once — no subscription.
Get the kit — $59One zip: the five-service compose stack with pinned versions, the security and retention layers raw Loki lacks, four provisioned dashboards, and the playbook library — plain files you own, on your own server.
Vector + Loki + Grafana + Alertmanager + Caddy in one docker compose project — official upstream images pinned by tag, read-only mounts, and a setup.sh that generates secrets, wires everything and verifies itself. From a fresh VPS to your first log line in about 30 minutes.
Grafana Loki ships without an authentication layer — their own install docs say so. The kit fronts the Loki API with Caddy TLS and bcrypt basic auth generated at install, so your logs are never the open port on the box. Grafana keeps its own login; signup is disabled.
Three retention formulas with sizing notes, enforced by Loki's compactor — plus a host-level disk-watch that alerts your webhook when usage crosses 80%. The quiet full disk is the most common way self-hosted Loki dies; this kit closes that door.
Three shipped rules — high error rate, pipeline stalled, container logs missing — evaluated by Loki's ruler and delivered to Telegram or Slack by Alertmanager. setup.sh writes the config from your bot token or webhook, and the docs include the delivery test most setups skip.
Logs overview, NGINX access, services and apps, system journal — provisioned automatically on first boot with the Loki datasource. A demo heartbeat keeps Grafana non-empty until your real sources land, then it deletes cleanly.
A 24-query LogQL cookbook with every line explained, ingestion playbooks for NGINX, Node.js, Python and systemd/journald, the capacity planner (GB/day to disk/RAM) and the upgrade runbook — one service at a time, verify between steps, 30-second rollback.
No APM. This is logs: aggregation, search, dashboards and alerts. Metrics, traces and profiling are different disciplines — pairing this kit with a metrics stack is your call, not a bundled promise.
No infinite scale. The shipped config targets the 1–5 GB/day class on a single VPS — the honest range for most small teams. The capacity planner states that ceiling and the escalation ladder for when you outgrow it: trim ingestion, grow the disk, move storage to object storage.
No managed service. You own the box, and upgrades are your 15 minutes a month. The runbook makes them boring — pinned versions, one service at a time, verify between steps — but nobody does them for you.
From zip to a working log stack on day one — then about 15 minutes a month.
One command generates secrets (chmod 600), picks 7/30/90-day retention, wires your Telegram or Slack channel and brings up all five services. tools/verify.sh then checks Grafana, Loki auth, log flow and alert rules — its exit code fits cron.
Every container's stdout is collected and JSON-parsed out of the box. NGINX, Node.js, Python and systemd each have a copy-paste playbook — with the canary command that proves ingestion actually works.
disk-watch.sh on the host cron, verify.sh weekly, a capacity check monthly, upgrades by the runbook. That is the entire maintenance loop — and your log bill stays whatever your VPS costs.
Published pricing of the log-management category, checked 2026-09-20. Subscriptions change often — verify before you buy anything.
| Option | Published price | What is metered | The catch |
|---|---|---|---|
| Datadog | $0.10/GB ingest + $1.70 per million events indexed (15-day retention, annual rate) | every layer: ingest, indexing, retention, forwarding | usage-based bills spike; public bill stories include $65M in a quarter |
| Better Stack | Nano $25/mo (annual, EU) for 40 GB → Tera $420/mo for 700 GB | monthly subscription per bundle | predictable — but recurring forever, and your data sits with the vendor |
| Graylog | Open: free, unlimited volume (SSPL) → Enterprise from $15,000/year | nothing — until you need support or compliance, then a big jump | self-host possible; assemble auth, retention and alerting yourself |
| Grafana Cloud | Free: 50 GB/month, 14-day retention → Pro from $19/mo platform fee + per-GB charges | platform fee + process + write + retain, per GB | genuinely good free tier — but 14-day retention and vendor-held data |
| Log Ops Kit | $59 once + your own ~$6/mo VPS | nothing — flat files, your hardware, your disk | your logs stay on your server; you run the 15-minute maintenance loop |
Competitor figures are their published prices, checked 2026-09-20 (sources: datadoghq.com/pricing, betterstack.com/pricing, graylog.org/pricing, grafana.com/pricing); subscriptions change often, check before you buy anything. The kit is a one-time purchase — your recurring cost is exactly your VPS.
No per-GB meter, no per-seat plan, no retention upsell — the platforms already charge enough of those.
≈ ฿2,090
Secure checkout via Stripe · card payments · taxes calculated at checkout
No — it is a one-time purchase of files. The software inside (Vector, Loki, Grafana) is free and open source; what you pay for is the assembly: the auth layer Loki lacks, retention that actually deletes, the alert pipeline, the dashboards, and the documentation that turns six-to-ten hours of integration into about thirty minutes.
32 files: the compose stack with pinned image versions, setup.sh with its self-tests, the Caddyfile with the auth layer, three retention formulas, three alert rules, four provisioned Grafana dashboards, the 24-query LogQL cookbook, four ingestion playbooks (NGINX, Node.js, Python, systemd/journald), the capacity planner, the upgrade runbook, and three tools — verify, capacity, disk-watch.
You need Docker installed — one copy-paste line if the VPS is fresh. After that, bash setup.sh does the assembly and tools/verify.sh tells you if anything is wrong. The playbooks are written for people who can edit a config file, not for container engineers.
Your VPS — typically $6–12/month for the 1–5 GB/day class the kit targets (the sizing table in the capacity planner shows the math). No other recurring cost. For comparison, published SaaS pricing for the same ingest class starts around $25/month and rises with volume.
Because Grafana Loki ships without any authentication layer — that is documented, not a rumor. Every DIY setup that ends at "Loki exposed on port 3100" leaves your logs readable to the internet. This kit fronts Loki with TLS and bcrypt basic auth generated at install: closed from the first boot, not a follow-up task.
If the kit doesn't fit how you work, write to us within 14 days of purchase and we refund in full — the files are yours to keep or delete. One refund per purchase; we only ask for optional feedback.
$59 one-time · ≈ ฿2,090 · free updates v1.x · instant download
Get the Log Ops Kit — $59