Question
· 14 hr ago

CSPGatewayLatency Alert, what does it mean? How should be managed?

Hello, thanks for your time reading this question.

We are receiving each day, alerts from one of our four Production nodes. It always has the same text:

 

[InterSystems IRIS SEVERE ERROR gchciris4.canariasalud:ENSEMBLE] [Utility.Event] [SYSTEM MONITOR] CSPGatewayLatency(127.0.0.1:443) Alert: CSPGatewayLatency = 5001.304, 5001.233, 5000.964 (Max value is 2000).

 

We have looked for it in the documentation here:

https://docs.intersystems.com/latest/csp/docbook/DocBook.UI.Page.cls?KEY...

 

We thought it was related to the following paragraph:

"Server Response Timeout

The maximum number of seconds allowed for the target Caché server to respond to a request from the web server. The timeout refers to a period of no activity, so, for example, sending a line of HTML data every second for 10 hours does not cause a timeout. The minimum allowable value for this field is 5 seconds.

The value set here is the default for the system. If an Inherited Value is specified, the value came from the Default Parameters page. You may, however, set a different value on individual server-specific configurations or within the application itself.

Note that if you have an Apache server, you can also set this value using Timeout in the Apache http.conf file. The lower of these two values is triggered first."

 

 

However, we suppose that CSP Gateway Latency of maximum 2 seconds is not related to Server Response Timeout by default limited to 5 seconds.

 

We just want to understand the Node error. Could you please point to the documentation where it is explained?

 

We need to understand it prior to solve it.

 

We have also looke for "CSPGatewayLatency" using the forum search text area, but it does not pop up relevant results.

https://community.intersystems.com/smartsearch?search=CSPGatewayLatency

 

Could you please help us? Thank you so much.

 

We have also seen this example:

Update Default Parameter(s)

Kill newpars
Set newpars("Server_Response_Timeout")=30
Do gateway.SetDefaultParams(.newpars)

Found here: https://docs.intersystems.com/latest/csp/docbook/DocBook.UI.Page.cls?KEY...

 

And this old post:

https://community.intersystems.com/post/cache-monlbl-response-time-question

 

Could your help us, please? Thanks.

Product version: IRIS 2020.1
$ZV: IRIS for UNIX (Red Hat Enterprise Linux for x86-64) 2020.1.1 (Build 408U) Sun Mar 21 2021 22:21:14 EDT
Discussion (2)2
Log in or sign up to continue

The CSPGatewayLatency alert indicates that the latency experienced in the CSP gateway exceeded the predefined maximum value. Here are the details:

  1. Understanding CSPGatewayLatency Alert:
    CSPGatewayLatency measures the delay involved in processing requests between the CSP gateway and your web server. The alert you received indicates that the latency has far surpassed the maximum allowable value (2000 ms), reaching values exceeding 5000 ms. This suggests a significant performance issue potentially caused by a delay in processing requests made to the web server or server-side resources [1][2].

  2. Relation to Server Response Timeout:
    While your suspicion linking CSPGatewayLatency to server response timeout is reasonable, these are distinct metrics. The Server Response Timeout is explicitly the maximum allowed time for the target InterSystems IRIS or Caché server to respond to incoming requests. On the other hand, CSPGatewayLatency is broader, detecting higher latency in communication specifically at the gateway level, which could result from multiple latency factors, not just server response time exceeding its timeout [1].

  3. Investigating and Managing This Alert:
    a. Check logs (e.g., WebGateway Event Log) for more in-depth details about what requests might be causing increased latency [3].
    b. Examine server-side performance metrics to identify potential bottlenecks—this could involve queries or processes that exceeded the Server Response Timeout set in your configurations [1][2].
    c. Assess current CSP Gateway configuration, such as limits on server connections (max connections restrictions) and adjust as needed [2].

Upon troubleshooting, you may identify that specific CSP requests or load on your server resulted in slow response times. You can address server-side bottlenecks, optimize database queries or increase relevant timeout thresholds cautiously.

Sources: