Datadog
The value is the network, the data or the infrastructure. Keep paying.
An observability platform: agents collect metrics, traces and logs from everything you run, and the product correlates them, alerts on them and keeps them for as long as you pay to.
Do not rebuild Datadog. Assemble the open stack, know its limits, and keep the bill visible. WHAT TO RUN - Prometheus for metrics, with remote write into Thanos or Mimir if retention beyond 15 days is needed - Grafana for dashboards - Loki for logs, Tempo for traces, both fed by the OpenTelemetry Collector - Alertmanager for routing THE COLLECTOR IS THE CONTRACT - One OpenTelemetry Collector config, checked in, deployed everywhere - Every service exports OTLP and nothing else, so a backend can be swapped without touching application code CARDINALITY, WHICH IS WHAT KILLS THIS - Ban unbounded labels: user id, request id, full URL path - A metric relabel config that drops anything not on an allow-list - An alert on series count per job, before the storage fills ALERTS WORTH HAVING - Symptom alerts on latency, error rate and saturation, not on CPU - Every alert carries a runbook link, and one without a link fails review RETENTION AND COST - 15 days of raw metrics, 90 days downsampled, 7 days of logs, 3 days of traces - Write the monthly storage cost of each into the README, and the person who owns the bill WHAT YOU ARE TAKING ON - Upgrades, capacity, and being paged when the monitoring itself is down Include a runbook for the day the metrics stop arriving.
What you lose
- Ingest and retention of metrics, traces and logs at a volume that would otherwise be a team of its own
- Correlation across all three from one page, which is what shortens an incident
- Hundreds of maintained integrations that work the day you install the agent
- Alerting with anomaly and forecast detection rather than fixed thresholds
- Someone else being on call for the monitoring system itself
If you would rather not build
- New Relic — paid, similar scope, different pricing model
The escape hatch
open source · no votes, no paid placement
SigNoz
$0Metrics, traces and logs in one self-hosted product, built on OpenTelemetry.
SigNoz/signozfree · open source
Grafana
$0Dashboards over Prometheus, Loki and Tempo — the front end of the open stack.
grafana/grafanafree · open source
Why this verdict
our own opinion · changed only by a person
14/100
Verdict no at 14: the open stack is genuinely good and this prompt assembles it properly, but you are hiring yourself as the observability team. Datadog is expensive because ingest at scale is expensive, and the correlation across signals is what actually ends an outage.
History
tracked since 9 Aug 2026 · nothing is ever overwritten
Questions about Datadog
answered from the record above
Is Datadog free?
No — the plan we track is $15 a month. Pro at $15 per host per month billed annually, $18 monthly, before logs, APM, synthetics and custom metrics — which is where the bill actually comes from.
Can you replace Datadog by building your own?
KEEP IT. The value is the network, the data or the infrastructure. Keep paying. Replacement score 14 out of 100, build time longer than it saves. Read what you lose before you decide.
How much does Datadog cost?
$15 a month on Pro — $180 a year. Recorded 9 Aug 2026.
What do you lose by replacing Datadog?
Ingest and retention of metrics, traces and logs at a volume that would otherwise be a team of its own; Correlation across all three from one page, which is what shortens an incident; Hundreds of maintained integrations that work the day you install the agent; Alerting with anomaly and forecast detection rather than fixed thresholds; Someone else being on call for the monitoring system itself. If any of those carry weight for you, keep paying.
Is there an open-source alternative to Datadog?
Yes: SigNoz, Grafana. The prompt on this page is for when you want it your way instead.
Related entries
same category first, most replaced first
Visualping
UptimeRobotEvery week, something stops being worth paying for.
New verdicts, prices that moved, entries added. One email a week. Unsubscribe in one click.
free forever · no tracking pixel · every prompt stays free on the site