Skip to main content
July 9, 2026
Question

Chat Agent – Tool Execution Failure ("Downstream tool request failed")

  • July 9, 2026
  • 4 replies
  • 0 views

Hi Team,

Our chat agent is intermittently failing when attempting to execute tool calls. End users are receiving one of the following error messages during these interactions:

  • Tool execution failed: Downstream tool request failed
  • I'm currently experiencing a system issue retrieving the data.

4 replies

July 9, 2026

Review Appian's integration logs (tomcat-stdOut.log or the specific integration error logs in the Monitoring view) for HTTP 429 Too Many Requests errors that correlate with the exact timestamps.

July 9, 2026

Hi Kapil,

When i checked the log, this is the error message we are getting.

OkHttp ai-platform.ai-platform.svc.cluster.local/...] ERROR com.appiancorp.lcpapi.streaming.StepStreamingEventSourceListener - Upstream SSE stream failed for run
java.net.SocketException: Socket closed
at java.base/sun.nio.ch.NioSocketImpl.ensureOpenAndConnected(NioSocketImpl.java:165)
at java.base/sun.nio.ch.NioSocketImpl.beginRead(NioSocketImpl.java:236)
at java.base/sun.nio.ch.NioSocketImpl.implRead(NioSocketImpl.java:304)
at java.base/sun.nio.ch.NioSocketImpl.read(NioSocketImpl.java:355)
at java.base/sun.nio.ch.NioSocketImpl$1.read(NioSocketImpl.java:808)
at java.base/java.net.Socket$SocketInputStream.read(Socket.java:966)
at okio.InputStreamSource.read(JvmOkio.kt:93)
at okio.AsyncTimeout$source$1.read(AsyncTimeout.kt:128)
at okio.RealBufferedSource.request(RealBufferedSource.kt:209)
at okio.RealBufferedSource.require(RealBufferedSource.kt:202)
at okio.RealBufferedSource.readHexadecimalUnsignedLong(RealBufferedSource.kt:384)
at okhttp3.internal.http1.Http1ExchangeCodec$ChunkedSource.readChunkSize(Http1ExchangeCodec.kt:437)
at okhttp3.internal.http1.Http1ExchangeCodec$ChunkedSource.read(Http1ExchangeCodec.kt:416)
at okhttp3.internal.connection.Exchange$ResponseBodySource.read(Exchange.kt:281)
at okio.RealBufferedSource.select(RealBufferedSource.kt:232)
at okhttp3.internal.sse.ServerSentEventReader.processNextEvent(ServerSentEventReader.kt:50)
at okhttp3.internal.sse.RealEventSource.processResponse(RealEventSource.kt:75)
at okhttp3.internal.sse.RealEventSource.onResponse(RealEventSource.kt:46)
at okhttp3.internal.connection.RealCall$AsyncCall.run(RealCall.kt:519)
at java.base/java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136)
at java.base/java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635)
at java.base/java.lang.Thread.run(Thread.java:840)

July 9, 2026

Hi [mention:81ab632089fd4d619d6816189ea9c949:e9ed411860ed4f2ba0265705b8793d05] ,

Server-Sent Events (SSE) require a long-lived, persistent HTTP connection to stream the LLM's response back to the user. If your downstream tool takes 30 to 60 seconds to execute, the AI pauses streaming.
I would suggest check the idle timeout and read timeout settings of your downstream service.

Moreover, the Appian chat component is waiting for the AI, and the AI is waiting for your tool to finish its work. If the tool hangs or takes too long, the underlying TCP socket simply gives up and drops.

Can you try increasing these timeouts to accommodate the maximum duration your longest tool takes to run, plus the standard LLM processing time.