Uptime Monitoring Dashboard for HTTP Endpoint Health

Use this dashboard to monitor synthetic HTTP checks across availability, response content validation, phase-level latency, payload size, and TLS certificate expiry.

The OpenTelemetry Collector's HTTP Check Receiver probes each endpoint on a schedule and reports the result as metrics. This dashboard reads those httpcheck.* metrics.

Dashboard Coverage

Use these panels to:

• Answer "is it up": See the overall 2xx rate, how many endpoints are reporting, how many are failing outright, and the mean probe duration.
• Find which endpoint broke: Compare per-endpoint availability over time and read the exact failure text for endpoints that never got a response.
• Catch a 200 that lies: Assert on response content, so an endpoint that returns HTTP 200 with an error body is flagged instead of counted as healthy.
• Locate the slow phase: Split probe time into DNS, TCP connect, TLS handshake, request write, and response read.
• Stay ahead of certificate expiry: Track time remaining per certificate, with issuer and common name.
• Inspect status codes: Break probes down by endpoint, status code, and status class.

Required Collector Metrics

The receiver enables only httpcheck.status, httpcheck.duration, and httpcheck.error by default. Every other metric that this dashboard reads is off until you enable it.

Step 3 of Monitor HTTP Endpoints enables everything except the two validation metrics, which need per-target assertions as well.

Metrics Included

Availability

• Availability: Share of probes that returned a 2xx response (httpcheck.status).
• Endpoints Monitored: Number of endpoints reporting probe results.
• Endpoints In Error: Endpoints whose probe failed before any HTTP response (httpcheck.error).
• Avg Response Time: Mean total probe duration (httpcheck.duration).
• Availability by Endpoint: Per-endpoint 2xx rate over time.
• Transport Errors by Endpoint: Probes that never received a response.
• Status Class Distribution: Share of probe results per status class.
• Transport Error Messages: The failure text per endpoint (error.message), with a count of affected probes. This is the reason behind the error count: refused connections, DNS failures, TLS verification errors.
• Endpoint Availability: 2xx rate per endpoint over the window, sortable worst-first.

Content Validation

Reachability alone cannot tell you whether a response is correct. These panels read httpcheck.validation.passed and httpcheck.validation.failed, which the receiver emits when you configure assertions on the response body.

• Validation Pass Rate: Share of content assertions that passed.
• Validation Failures: Total failed assertions in the window.
• Validation Failures by Endpoint and Check: Failures over time, split by endpoint and validation.type.
• Failed Content Checks: Which assertion failed, on which endpoint, and how often. Cross-check this against Status Codes by Endpoint: a row here whose status code is 2xx is a silent failure.

Latency

• Total Response Time by Endpoint: Full probe wall time, DNS through response body (httpcheck.duration).
• DNS Lookup Duration: Time spent resolving the hostname.
• TCP Connect Duration: Time to establish the transport connection.
• TLS Handshake Duration: Time spent on the TLS handshake.
• Request Write Duration: Time spent writing the request.
• Response Read Duration: Time spent receiving the response.
• Response Time by Endpoint: Mean probe duration per endpoint, slowest first.

Payload & TLS

• Soonest Cert Expiry: Least time remaining on any verified certificate, with a threshold that turns red inside 14 days.
• Avg Response Size and Max Response Size: Response body size across the selected endpoints.
• Response Size by Endpoint: Body size per endpoint. A sudden drop often means an error page replaced the real response.
• TLS Cert Time Remaining: Seconds until expiry per endpoint.
• TLS Certificates: Time remaining per certificate with issuer, common name, and SAN, soonest expiry first.

HTTP Status Detail

• Status Codes by Endpoint: Probe count per status code each endpoint returned.
• Non-2xx Responses by Endpoint: Probes that got a response outside 2xx, split by endpoint and status class.

Dashboard Variables

Use these filter variables:

• Service: Filter by the Collector service emitting the metrics (service.name).
• Endpoint: Filter by probed endpoint URL (http.url).

Panels That Read Empty

Some panels stay empty until the matching data exists, which is expected rather than a misconfiguration:

• Any optional-metric panel whose metric is still disabled in the Collector. See Required Collector Metrics.
• Content Validation panels need validations configured on your targets. Enabling the two metrics alone produces nothing.
• Response size panels need a response body of at least one byte. An endpoint that returns 200 with an empty body reports no size.
• TLS panels only report for endpoints whose certificate verifies. An expired or untrusted certificate surfaces as a transport error instead, because the client refuses the handshake.
• DNS Lookup Duration and TLS Handshake Duration filter out zero values, so endpoints that resolve without a lookup or reuse a connection do not appear.

Next Steps

• Alert on availability and certificate expiry from a panel in this dashboard.
• Add response content validation to fill the Content Validation panels.
• Set up a notification channel for Slack, PagerDuty, or email.

Related Dashboards

• APM Metrics
• HTTP API Monitoring
• Cert Manager