What causes "Already closed {}" in pinpoint-apm/pinpoint?

Remove redundant close() calls from application/plugin code and let the agent manage the trace lifecycle

Already closed

Error message

Already closed {}

What it means

ChildTrace.close0() logs 'Already closed {}' as a warning when close0() is invoked on a ChildTrace instance whose `closed` flag is already true. ChildTrace is a lightweight per-span trace object, and the library guards against double-close so the underlying TraceRoot is not finalized twice. The warning means some code path closed the trace more than once; the second close is a no-op.

Solutions

Remove redundant close() calls from application/plugin code and let the agent manage the trace lifecycle
Guard your own cleanup with an idempotency check (e.g. a local AtomicBoolean) before calling close()
Ensure each Trace obtained from TraceContext is closed exactly once per request thread
If seen after upgrading, check the agent version for known double-close bugs in async trace handling

Example fix

// before
Trace trace = traceContext.newTraceObject();
doWork();
trace.close();
trace.close(); // warns Already closed
// after
Trace trace = traceContext.newTraceObject();
doWork();
trace.close();

Defensive patterns

Strategy: try-catch

Validate before calling

// check before cleanup
if (trace != null && !isTraceClosed(trace)) {
    trace.close();
}

Type guard

boolean isOpen(Trace t) { return t != null && !isClosedFlagSet(t); }

Try / catch

try { trace.close(); } catch (RuntimeException e) { logger.debug("trace close skipped", e); }

Prevention

Give one component sole ownership of trace close()
Use try/finally so close happens exactly once per request
Never reuse a Trace object after close()
Log at debug level in your own double-close guards

When it happens

Trigger: Calling close() (which delegates to close0()) twice on the same ChildTrace, e.g. manual close() followed by framework/agent-triggered cleanup, or a finally-block close after the agent already closed the trace when the call stack was force-completed.

Common situations: Plugin/interceptor code that explicitly closes traces plus an agent-side cleanup path both firing; async trace completion racing with synchronous close; reusing a Trace reference after its lifecycle ended.

Understand the failure class

Background: "Invalid state transition" errors: "status must be X, actually Y", "already rejected/charging/uninstalled", "cannot ... while running" — what they mean when a library rejects your call — this error's family across 31 libraries.

Related errors

AI-assisted analysis of pinpoint-apm/pinpoint@744c3d3075 (2026-09-07). Data as JSON: /api/errors/e68a4ad62d32034f. Report an issue: GitHub.

Appendix: source

Thrown at agent-module/profiler/src/main/java/com/navercorp/pinpoint/profiler/context/ChildTrace.java:177

    }


    @Override
    public boolean isClosed() {
        return closed;
    }

    @Override
    public void close() {
        if (asyncTraceBlock) {
            traceBlockEnd(ASYNC_BEGIN_STACK_ID);
        }
        close0();
    }

    public void close0() {
        if (closed) {
            logger.warn("Already closed {}", this);
            return;
        }
        closed = true;

        if (!callStack.empty()) {
            if (logger.isWarnEnabled()) {
                stackDump("not empty call stack");
            }
            // skip
        } else {
            logSpan();
        }
        this.wrappedSpanEventRecorder.close();
        this.storage.close();
    }

    @Override
    public TraceId getTraceId() {

View on GitHub (pinned to 744c3d3075)