Response Flags in Access Logs
UH
Description
UH (NoHealthyUpstream) indicates that there are no healthy service instances at the backend.
Typical Symptom
All service instances at the backend are unavailable. For example, the number of service instances is set to 0.
Example Log
Client log
Suggestion
Check the pod configuration of the target service and ensure that all service instances are running properly.
DC
Description
DC (DownstreamConnectionTermination) indicates that the downstream connection is terminated.
If the client disconnects from the server before receiving a response due to various reasons, DC will be recorded in the access logs. In most cases, the size of the response body is 0.
Typical Symptom
When accessing the target service, the client disconnects from the server before receiving a response or receives no response from the server within the timeout set on the client.
Example Log
Client log
Suggestion
Generally, no operation is required. In most cases, time-consuming responses often result in client disconnections. You can optimize the target service to solve such an issue.
UF
Description
UF (UpstreamConnectionFailure) indicates an upstream connection error.
Typical Symptom
The service port of the target service is unreachable. For example, if the client accesses the target service over an invalid port, 503UF will be recorded in the outbound logs of the client agent.
If the pod port of the target service is unreachable, 503UF will be recorded in the inbound log of the server.
Example Logs
Outbound log generated for an invalid service port
Inbound log generated for an invalid pod port
Suggestion
Specify a valid service port, so that the client accesses the target service over this port.
URX
Description
URX (UpstreamRetryLimitExceeded) indicates that the requests are rejected because the HTTP retry limit or the maximum number of TCP connection attempts is reached.
Typical Symptom
If the number of failures to invoke the server reaches the retry limit configured for the target service in the service mesh, or the failure persists after two retries (default value in Istio), URX will be recorded in the outbound logs on the client.
Example Log
Outbound log on the client
Suggestion
Improve the service success rate, so that services can be properly accessed after limited retries.
UPE
Description
UPE (UpstreamProtocolError) indicates an upstream protocol error.
Typical Symptom
If an application protocol defined for each service port in the service mesh is inconsistent with the actual protocol, the client will receive protocol error code 502 during service access. 502 UPE will be recorded in the inbound logs on the server.
Example Logs
Outbound log on the client
Inbound log on the server
Suggestion
Define a correct application protocol for each service port. Ensure that the protocol is read from the port name of the service in Istio, or use the AppProtocol field to specify a correct application protocol for each service port.
DPE
Description
DPE (DownstreamProtocolError) indicates a downstream protocol error.
Typical Symptom
The downstream client uses an incorrect protocol to access the target service.
Example Logs
Outbound log on the client
Inbound log on the server
Suggestion
Ensure that the client uses a correct protocol to access the target service.
NC
Description
NC (NoClusterFound) indicates that the cluster where the service instance in the service mesh resides was not found.
Typical Symptom
The backend defined in the traffic distribution policy is not defined in the service mesh. For example, as defined in the traffic distribution policy, traffic is routed to the service subset identified by V2, but this service subset is not defined in DestinationRule.
Example Log
Client log
Suggestion
Ensure that the service subset referenced in VirtualService is correctly defined in DestinationRule.
NR
Description
NR (NoRouteFound) indicates that there is no matched route for the requests. Generally, status code 404 is returned.
Typical Symptom
The requests do not meet the route matching conditions defined in VirtualService. As a result, no route is matched for the requests.
Example Log
Outbound log on the client
Suggestion
Ensure that client requests meet the traffic characteristics defined in the route, so that the route defined for the server is matched for the requests.
DI
Description
DI (DelayInjected) indicates that a delay fault is injected into the request.
Typical Symptom
If delay fault injection is configured in VirtualService, the request processing will be delayed for a period specified via fault injection, and DI will be recorded in the access logs.
Example Logs
Outbound log on the client
Inbound log on the server
Suggestion
No operation is required. This is a manual operation recorded in the access log.
FI
Description
FI (FaultInjected) indicates the request was aborted with a response code specified via fault injection.
Typical Symptom
A fault with a specific status code is injected to the target service through VirtualService.
Example Log
Outbound log on the client
Suggestion
No operation is required. This is a manual operation recorded in the access log.
UT
Description
UT (UpstreamRequestTimeout) indicates that an upstream request times out. Generally, HTTP status code 504 is returned.
Typical Symptom
A timeout is configured for the target service in VirtualService. When the specified time period expires, the client agent automatically times out and cancels the request.
Example Log
Outbound log on the client
Suggestion
No operation is required. The timeout is configured for quick access failure of the target service. If the service access times out, the response flag will be returned.
RL
Description
RL (RateLimited) indicates that the traffic limit is triggered. Generally, HTTP status code 429 is returned.
Typical Symptom
A local or global traffic limit policy is configured in the service mesh. If the number of requests in a unit time reaches the specific threshold, the traffic limit is triggered.
Example Logs
Traffic limit policies apply to the scenario where the server receives the traffic
Outbound log on the client
Inbound log on the server
Traffic limit policies apply to the scenario where the client sends the traffic
Suggestion
No operation is required. Traffic limit policies are configured to protect the target service.
Feedback
Was this page helpful?
Provide feedbackThank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.
For any further questions, feel free to contact us through the chatbot.
Chatbot