Integrations: Spitfire works with the tools you already run
A load test tells you where the system broke. The integrations add what happened inside at that moment: your monitoring's alerts, your OpenTelemetry traces and metrics. And the result goes where the team already works: Slack, Jira, the pipeline.
The run broke at the 450 VU stage (p95 900 ms, limit 500 ms). The alert closest to the break: Datadog, "DB connections high" on orders-db, opened 12 s before the break. At the same time orders-db's share of the time rose from 12% to 71% (40 traces).
Monitoring and APM alerts
Whatever your monitoring alerted on lands on the run, at the load stage where the system broke: which host, when, how many seconds before the break.
- ZabbixProblems opened in the run window, filtered by host group, host and severity. Zabbix 5.4 and later.
- DatadogMonitors that alerted and their recoveries; an Application key with only monitors_read and events_read.
- New RelicAlert incidents and their closings over NerdGraph; a User key is enough.
- DynatraceProblems opened and closed; an API token with only the problems.read scope.
- InstanaIssues and incidents opened during the run; an API token.
Monitoring events add-on: included in Enterprise; a 7-day trial on every other edition, the free one too. Read-only access, nothing is written to these systems. Zabbix, Datadog, New Relic, Dynatrace or Instana also come one by one: $13 / month or $129 / year each. Pick 3 or more and you get the whole add-on for the same price. Pricing →
OpenTelemetry and open observability
No agent: the traces, metrics and logs your systems already produce are aligned with the load stages; the Backend tab writes which service, database and resource changed by how much.
- OpenTelemetry (OTLP)One exporter on your Collector: traces, metrics, error and warning logs also reach Spitfire during the run.
- PrometheusYour own PromQL: p95, error ratio, CPU, memory and connection pools per service. Grafana Mimir, Thanos and VictoriaMetrics too.
- Tempo · JaegerEach stage's traces; the slowest request's spans open as a waterfall.
- LokiError lines per service (LogQL), stage by stage.
In every edition, the free one included.
Issue tracking
A run that failed a threshold, a worse comparison, a monitor going down or a schedule that did not start lands in your team's queue. A repeat is a comment, not a new issue; so is the recovery.
- JiraAn issue in the project you pick; Jira Cloud and Data Center.
- ServiceNowAn incident, with the assignment group, category, urgency and impact you pick. Repeats become work notes.
Issue tracking add-on: included in Enterprise; a 7-day trial on the other editions. Jira or ServiceNow also come one by one: $19 / month or $190 / year each. Pick them all and you get the whole add-on for the same price. Pricing →
Notifications
A run finished, a threshold broke, the new release got worse, a production journey went down: the message goes where the team looks, the PDF report is attached to the e-mail.
- SlackIncoming Webhook: the result, p95, error rate, failed thresholds and a link to the run.
- Microsoft TeamsAn adaptive card through a Workflows webhook.
- E-postaFrom your own SMTP server; the finished run's PDF report attached.
- SMSAny SMS provider's HTTP API or your own gateway; on a network without internet too.
- WebhookSigned JSON (HMAC-SHA256), retries and a delivery log.
In every edition; 1 channel on the free one.
CI/CD
Run a saved test from the pipeline; a broken threshold fails the step. Each step's p95, p99 and error rate is written as a comment on the pull or merge request.
- GitHub ActionsOne step, and pull request comments.
- GitLab CIOne job, and merge request comments.
- Jenkins · Azure PipelinesOne CLI command; it passes or fails by exit code.
In every edition; PR/MR comments on 1 repository on the free one.
Single sign-on
People sign in with their company account and their groups map to roles; a leaver loses access through your identity provider.
- OpenID ConnectEntra ID, Okta, Google, Keycloak…
- SAML 2.0ADFS, Entra ID, Okta, Keycloak, Shibboleth…
- LDAP · Active DirectorySign-in with the directory user name or e-mail.
- SCIM 2.0Provisioning and deprovisioning from Okta, Entra ID, OneLogin, JumpCloud.
On paid plans; SCIM on Enterprise.
Send run data out
So your monitoring can tell a load test from an incident: run start, end and live metrics flow into your own dashboards.
- OTLP pushVUs, request rate, error rate, p50/p95/p99, per step, live.
- Prometheus /metricsScraped with a personal API token.
- SIEM (syslog · webhook)The audit log streams chained and signed.
In every edition; SIEM streaming on Enterprise.
Frequently asked
Do I need to install an agent?
Where are the connection credentials kept?
Does an alert mean it caused the break?
My system is not on the list.
Spitfire installs on Docker or Kubernetes with one command; every testing feature and protocol is open in the free edition.