Search API
This is the main endpoint for executing queries in Humio. This endpoint streams results as soon as they are calculated, but depending on the query type filter or aggregate, the time of delivery changes. The following table illustrates this
| Query Type | Live Query | Standard Query |
|---|---|---|
| Filter | Streaming | Streaming |
| Aggregate | Error — use query jobs | Streaming — result only at end |
The endpoint streams results for filter queries as they happen.
For aggregate standard queries, the result is not ready until the query has processed all events in the query interval. The request is blocked until the result is ready. It is at this point that Humio sends the result back.
For aggregate live queries, this endpoint returns an error. What you want in this situation is to get a snapshot of the complete result set at certain points in time (fx every second), but the query end point does not support this behavior. Instead, you should use the query job endpoint and then poll the result when you need it. For more information on live queries, see polling query endpoint.
Request
To start a query, POST the query to:
POST /api/v1/repositories/$REPOSITORY_NAME/queryThe JSON request body has the following attributes:
| Name | Type | Required | Description |
|---|---|---|---|
queryString
| string | Yes | The actual query. See Query Language for details. |
start
| Time | No | The start date and time. This parameter tells Humio not to return results from before this date and time. See how to specify a time. |
end
| Time | No | The end date and time. This parameter tells Humio not to return results from after this date and time. See how to specify a time. |
isLive
| boolean | No |
Sets whether this query is live. Defaults to
false. Live queries are continuously updated.
|
timeZoneOffsetMinutes
| number | No |
Set the time zone offset used for bucket()
and timeChart() time slices, which is
significant if the corresponding span is
multiples of days. Defaults to 0 (UTC);
positive numbers are to the east of UTC, so for
UTC+01:00 timezone the value
60 should be passed.
|
arguments
| object | No |
Dictionary of arguments specified in queries with
?param or
?{param=defaultValue} syntax. Provided
arguments must be a simple dictionary of string values. If an
argument is given explicitly as in
?query(param=value) then that value overrides
values provided here.
|
If you use this API from a browser application, you may want to trigger
direct download. You can achieve this by adding the HTTP header
X-Desired-Filename to the request. That will result
in the response having the header Content-Disposition with the value
attachment; filename=\"DESIRED_FILE_NAME\".
Time Specification
There are two ways of specifying the start and
end time for a query:
Absolute Time
With absolute time, you specify a number that expresses the precise time in milliseconds since the Unix epoch (Unix time) in the UTC/Zulu time zone. This method is shown in the following example:
{
"queryString": "",
"start": 1473449370018,
"end": 1473535816755
}Relative Time
With relative time, you specify the start and end time as a relative
time such as 1minute or 24hours.
Humio supports this using relative time modifiers. Humio treats the
start and end times as relative times if you specify them as strings.
When providing a timestamp, relative time modifiers are specified relative to "now".
See the Relative Time Syntax reference page.
Note
Relative time modifiers are always relative to now.
This method is shown in the following examples:
Search the last 24 hours:
{
"queryString": "ERROR",
"start": "24hours",
"end": "now"
}You can also mix relative and absolute time modifiers. For example, to search from a specified moment in time until two days ago:
{
"queryString": "loglevel=ERROR",
"start": 1473449370018,
"end": "2days"
}Note
Omitted and required arguments: Humio has defined behavior when you omit time arguments: if you omit the`end` argument, it gets the default value now; and if you omit the start argument, it gets the default value of 24hours. For *_live queries_*, you must either set end to "now", or omit it. You must set start to a relative time modifier.
Response
Humio returns data in different formats depending on the media type you set in the ACCEPT header of the HTTP request.
Data can be returned in the following formats:
| Media Type | Description |
|---|---|
text/plain (default)
|
Returns events delimited by newlines. If the event has a
rawstring field, then Humio prints it as
the event. If it does not, then Humio prints all fields on the
event in the format a->hello,
b->world. Note that the event can contain
newlines. Nothing is escaped.
|
application/json
|
Returns events in a standard JSON array. All field values in
each event are returned as JSON strings, except for
@timestamp. The
@timestamp field is returned a long
integer, representing time as Unix time in milliseconds
(UTC/Zulu time). Newlines inside the JSON data are escaped as
\n.
|
| application/x-ndjson | Returns events as Newline Delimited JSON (NDJSON). This format supports streaming JSON data. Data is returned with one event per line. Newlines inside the JSON data are escaped as n. |
Live Query Streaming All Events
This live query returns an empty search, finding all events in a time window going 10 seconds back in time.
Notice the ACCEPT header. This tells the server to
stream data as
Newline
Delimited JSON.
curl https://$YOUR_HUMIO_URL/api/v1/repositories/$REPOSITORY_NAME/query \
-X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H 'Content-Type: application/json' \
-H "Accept: application/x-ndjson" \
-d '{"queryString":"","start":"10s","isLive":true}'Aggregate Query Returning Standard JSON
This query groups results by service and counts the number of events for each service. The query blocks until it is complete and returns events as a JSON array:
curl https://$YOUR_HUMIO_URL/api/v1/repositories/$REPOSITORY_NAME/query \
-X POST \
-H "Authorization: Bearer $API_TOKEN" \
-H 'Content-Type: application/json' \
-H "Accept: application/json" \
-d '{"queryString":"count()","start": "1h","end":"now","isLive":false}'Query Jobs
Create
The Query Jobs endpoint lets you run a query and check its status later.
To execute a query using the Query Jobs endpoint, you first have to start it, and then subsequently poll its current status in a separate request.
The Query Jobs endpoint is designed to support the web front end. This means that filter queries only return a maximum of 200 matching events and aggregate queries up to a maximum of 1500 rows. The API has facilities to support Web UIs of the Query Jobs poll endpoint).
Note
You can use add the tail() function as a filter
to return more than 200 rows. For example "localhost | tail(10000)"
will return up to 10000 rows matching localhost.
Request
To start a Query Job, POST the query to:
POST /api/v1/repositories/$REPOSITORY_NAME/queryjobsThe request body is similar to the in the query endpoint.
Response
Starting a query yields a response of the form:
HTTP/1.1 200 OK
{ id: "some-long-id" }
The id field indicates the $ID
for the query, which you can then poll using the HTTP GET method (see
Poll).
Example
curl https://$YOUR_HUMIO_URL/api/v1/repositories/$REPOSITORY_NAME/queryjobs \
-X POST \
-H 'Authorization: Bearer $API_TOKEN' \
-H 'Content-Type: application/json' \
-H 'Accept: application/json' \
-H 'Accept-Encoding: ' \
-d '{"queryString":"","start": "1d","showQueryEventDistribution":true,"isLive":false}'Poll
This endpoint lets you poll running Query Jobs.
Request
To poll a running Query Job, make an HTTP GET request to the job.
In the following example request, replace $ID with
the ID from the response of the Query Jobs:
GET /api/v1/repositories/$REPOSITORY_NAME/queryjobs/$IDResponse
When Humio runs a search, it returns partial results. It returns the results that it found at the time of the polling. Humio searches the newest data first, and then searches progressively backward in time.
This way, Humio produces some results right away. The done:
true property in a poll query shows if the query is
finished.
The response is a JSON object with the following attributes:
| Name | Type | Description |
|---|---|---|
done
| boolean | True if the query has run to completion. |
events
| Array[Event] | The 200 most recent elements of the query response. |
metaData
| QueryMetaData | Information about the query. |
queryEventDistribution
| EventDistData |
Information used to render the distribution graph. Only
present in the response if
showQueryEventDistribution was set to true.
|
The MetaData field contains the number of matching
events, the query boundary, and information about the attributes and
their unique value domains in the response.
Warning
Query timeouts: If you do not poll a query for 30 seconds, then it stops and deletes itself. Live queries keep running for an hour without being polled.
Example
curl https://$YOUR_HUMIO_URL/api/v1/repositories/$REPOSITORY_NAME/queryjobs/$ID \
-H "Authorization: Bearer $API_TOKEN"Delete
Stops running Query Jobs.
Request
To stop a Query Job, you can issue a DELETE request
to the URL of the Query Job:
DELETE /api/v1/repositories/$REPOSITORY_NAME/queryjobs/$IDResponse
Standard HTTP response codes.
Example
curl https://$YOUR_HUMIO_URL/api/v1/repositories/$REPOSITORY_NAME/queryjobs/$ID \
-X DELETE \
-H "Authorization: Bearer $API_TOKEN"