Select count measures how many times the server enters the select loop to check for ready events, which is a low level indicator of request handling activity. Session count measures how many sessions have been created. The first reflects event processing load, while the second reflects application session behavior and user interaction patterns.
Jetty select count vs session count: what each metric is actually telling you
Jetty select count is a transport and event-loop signal, while session count is an application-state signal. They describe different layers of the stack: one tells you how often Jetty is polling for ready I/O, the other tells you how many application sessions have been created and are being tracked. Because they live at different layers, they should never be compared as if they measured the same thing.
That distinction matters operationally. A rising select count can indicate more event-loop activity without saying anything direct about logins, carts, or user state. A rising session count can indicate more users or more retained state, but not necessarily more low-level socket or request dispatch activity. In other words, select count is about reactor-style event processing, session count is about application affinity and session lifecycle.
How to interpret the two counters without mixing layers
Use select count when you are looking for evidence of request-handling churn, selector wakeups, or how busy the server is at the I/O layer. It is useful for understanding whether the server is spending time entering the select loop and responding to ready events, but it is not a proxy for user population. Session count is the better signal when you need to understand how much server-side session state the application is holding.
Those signals can move independently. A high-traffic API endpoint may generate substantial select activity with little or no session creation. A stateful web application may accumulate sessions even when the underlying request path is relatively steady. The correct interpretation depends on whether you are diagnosing connection and event processing behaviour or application state growth.
- Jetty documentation is the right place to verify how a specific metric is exposed in your Jetty version and what object or subsystem publishes it.
- Jetty documentation hub helps confirm whether the count comes from a connector, selector, or session manager view.
How practitioners should use these metrics in troubleshooting
What to verify: Check whether the metric source is connector-level or session-manager-level before building an explanation from it. If select count rises, validate it against request rate, thread utilisation, and selector wakeup behaviour. If session count rises, verify it against authentication, stickiness, session timeout policy, and application design.
Common mistake: Treating select count as a user or session metric leads to bad diagnosis. A busy selector does not prove user growth, and a large session population does not prove I/O pressure. You need both signals, plus request and latency data, to separate transport load from stateful application behaviour.
What good looks like: The numbers are interpreted alongside their owning subsystem, and any sustained divergence is explained by the application pattern rather than assumed to be a Jetty fault. If session count grows without a matching business reason, the next question is whether sessions are being expired, invalidated, or retained longer than intended.
Practitioner takeaway: Read select count as an event-loop activity indicator and session count as a state-retention indicator; the useful diagnosis comes from matching each one to its own layer instead of blending them into a single notion of load.
Standards & Framework Alignment
This section maps relevant standards and security frameworks to the operational risks and controls described in this guidance.
CIS Controls v8 and NIST CSF 2.0 set the governance and control requirements practitioners need to meet.
| Framework | Control / Reference | Relevance |
|---|---|---|
| CIS Controls v8 | 5 — Account Management | Session count reflects application account/session state that must be managed and bounded. |
| 8 — Audit Log Management | Select count is a runtime activity signal that should be correlated with logs for diagnosis. | |
| Recommendation — Track and govern session-bearing accounts to keep state growth and access exposure under control. Correlate selector activity with logs to distinguish I/O churn from application-state growth. | ||
| NIST CSF 2.0 | DE.AE-2 — Anomalous Events are Detected | Unexpected changes in either counter can indicate anomalous runtime behaviour worth investigating. |
| PR.PT-1 — Audit/Log Records | Metric interpretation depends on telemetry that distinguishes event-loop activity from session lifecycle. | |
| Recommendation — Compare these counters to baseline behaviour to flag abnormal request or session patterns. Use complementary telemetry so select and session metrics can be interpreted in context. | ||
Related resources from NHI Mgmt Group
Deepen Your Knowledge
Reviewed and updated by the NHIMG editorial team on September 19, 2026.
NHI Mgmt Group — the #1 independent authority on Non-Human Identity, IAM, and Agentic AI security. nhimg.org
