Insights Request-Response Dashboard
LogScale logs all of its requests under the
#kind=requests. This is
similar to an access log except that it's in LogScale's format and it
logs all HTTP requests to the LogScale cluster. The logs for this can
also be found in the
humio-requests.log file. See how
our LogScale Internal Logging for LogScale works for more
Requests per Second
This a timechart showing the average number of HTTP requests made per second to the LogScale cluster across all nodes.
Response Times (Millis)
This time chart shows the 50th, 90th, 99th and 100th percentile of the response times. This is how long it takes for LogScale to respond to an HTTP request.
Each HTTP request will log an HTTP status code. This timechart shows the distribution of HTTP status codes greater than 400 per second over the last 24 hours across the LogScale cluster.
This timechart shows the average requests per second across all
LogScale nodes in the Ingest Endpoint: either the
Ingest endpoint Usage
This timechart shows again the average requests per second across all LogScale nodes in the Ingest Endpoint, but then splits it by the endpoint that has been accessed.
Ingest Requests With Statuscode >= 400
Each HTTP request will log an HTTP status code. This timechart in particular shows the distribution of HTTP status codes greater than 400 per second over the last 24 hours across the LogScale cluster that is trying to hit any of our Ingest endpoints. So this will show errors relating to the ingest.
Ingest Request Size (Bytes)
When you send ingest to LogScale via a log shipper, LogScale will log
details about the HTTP request containing the shipped logs. This
timechart shows the 50th, 75th and 99th percentile of the
contentLength size in bytes
within that HTTP request. This is useful to help identify if log
shippers are sending requests that are larger than expected.
See LogScale Limits for more information.
Ingest Request Responsetime
When LogScale receives an HTTP request to its ingest endpoint, it will log how long until LogScale receives it and sends a response back. This can be useful to identify if there is any lag between log shippers sending the data and the LogScale cluster responding. This timechart shows the 50th, 75th and 99th percentile of this reponse time in milliseconds.
Query Submits (Per Minute)
When a query is submited to LogScale, a node is responsible for co-ordinating that query. This timechart shows the number of query submits per minute for each host.
On a healthy system, one would expect this to be evenly distributed between LogScale nodes on which node receive this submission.
Query Submits Per Repository
This timechart shows the number of queries submitted on each repository in a cluster per minute.
This can be useful to identify the repositories that are receiving most of the query load in a system. If it's also a high volume datasource, top repositories on this graph might be a good canidate for [tagging] to increase query performance.
Streaming Query Submits
In LogScale, a streaming query can be defined as queries that only Transformation Queries and has no Aggregation Queries, or a non live query that uses aggregates. A live query that uses aggregate functions won't use the streaming queries Search API, but the Create.