mirror of
https://github.com/elastic/kibana.git
synced 2025-04-23 09:19:04 -04:00
This PR detects when claiming tasks fails due to inline scripts being disabled in Elasticsearch and improves the message reported in the Kibana log.
This commit is contained in:
parent
b61401233f
commit
d361a1ee9c
4 changed files with 279 additions and 12 deletions
167
x-pack/legacy/plugins/task_manager/lib/identify_es_error.test.ts
Normal file
167
x-pack/legacy/plugins/task_manager/lib/identify_es_error.test.ts
Normal file
|
@ -0,0 +1,167 @@
|
|||
/*
|
||||
* Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
|
||||
* or more contributor license agreements. Licensed under the Elastic License;
|
||||
* you may not use this file except in compliance with the Elastic License.
|
||||
*/
|
||||
import { identifyEsError, ESErrorCausedBy } from './identify_es_error';
|
||||
|
||||
describe('identifyEsError', () => {
|
||||
test('extracts messages from root cause', () => {
|
||||
expect(
|
||||
identifyEsError(
|
||||
generateESErrorWithResponse(
|
||||
[
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'root cause',
|
||||
},
|
||||
],
|
||||
{}
|
||||
)
|
||||
)
|
||||
).toContain('root cause');
|
||||
});
|
||||
|
||||
test('extracts messages from deep root cause', () => {
|
||||
expect(
|
||||
identifyEsError(
|
||||
generateESErrorWithResponse(
|
||||
[
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'root cause',
|
||||
},
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'deep root cause',
|
||||
},
|
||||
],
|
||||
{}
|
||||
)
|
||||
)
|
||||
).toContain('deep root cause');
|
||||
});
|
||||
|
||||
test('extracts messages from first caused by', () => {
|
||||
expect(
|
||||
identifyEsError(
|
||||
generateESErrorWithResponse(
|
||||
[
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'root cause',
|
||||
},
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'deep root cause',
|
||||
},
|
||||
],
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'first caused by',
|
||||
caused_by: {
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'second caused by',
|
||||
},
|
||||
}
|
||||
)
|
||||
)
|
||||
).toContain('first caused by');
|
||||
});
|
||||
|
||||
test('extracts messages from deep caused by', () => {
|
||||
expect(
|
||||
identifyEsError(
|
||||
generateESErrorWithResponse(
|
||||
[
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'root cause',
|
||||
},
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'deep root cause',
|
||||
},
|
||||
],
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'first caused by',
|
||||
caused_by: {
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'second caused by',
|
||||
},
|
||||
}
|
||||
)
|
||||
)
|
||||
).toContain('second caused by');
|
||||
});
|
||||
|
||||
test('extracts all messages in error', () => {
|
||||
expect(
|
||||
identifyEsError(
|
||||
generateESErrorWithResponse(
|
||||
[
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'root cause',
|
||||
},
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'deep root cause',
|
||||
},
|
||||
],
|
||||
{
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'first caused by',
|
||||
caused_by: {
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'second caused by',
|
||||
},
|
||||
}
|
||||
)
|
||||
)
|
||||
).toMatchInlineSnapshot(`
|
||||
Array [
|
||||
"first caused by",
|
||||
"second caused by",
|
||||
"root cause",
|
||||
"deep root cause",
|
||||
]
|
||||
`);
|
||||
});
|
||||
});
|
||||
|
||||
function generateESErrorWithResponse(
|
||||
rootCause: ESErrorCausedBy[] = [],
|
||||
causeBy: ESErrorCausedBy = {}
|
||||
) {
|
||||
return Object.assign(new Error(), {
|
||||
msg: '[illegal_argument_exception] cannot execute [inline] scripts',
|
||||
path: '/.kibana_task_manager/_update_by_query',
|
||||
query: {},
|
||||
body: '{"query":{}}',
|
||||
statusCode: 400,
|
||||
response: JSON.stringify({
|
||||
error: {
|
||||
root_cause: rootCause,
|
||||
type: 'search_phase_execution_exception',
|
||||
reason: 'all shards failed',
|
||||
phase: 'query',
|
||||
grouped: true,
|
||||
failed_shards: [
|
||||
{
|
||||
shard: 0,
|
||||
index: '.kibana_task_manager_1',
|
||||
node: '24A4QbjHSK6prvtopAKLKw',
|
||||
reason: {
|
||||
type: 'illegal_argument_exception',
|
||||
reason: 'cannot execute [inline] scripts',
|
||||
},
|
||||
},
|
||||
],
|
||||
caused_by: causeBy,
|
||||
},
|
||||
status: 400,
|
||||
}),
|
||||
});
|
||||
}
|
59
x-pack/legacy/plugins/task_manager/lib/identify_es_error.ts
Normal file
59
x-pack/legacy/plugins/task_manager/lib/identify_es_error.ts
Normal file
|
@ -0,0 +1,59 @@
|
|||
/*
|
||||
* Copyright Elasticsearch B.V. and/or licensed to Elasticsearch B.V. under one
|
||||
* or more contributor license agreements. Licensed under the Elastic License;
|
||||
* you may not use this file except in compliance with the Elastic License.
|
||||
*/
|
||||
|
||||
export interface ESErrorCausedBy {
|
||||
type?: string;
|
||||
reason?: string;
|
||||
caused_by?: ESErrorCausedBy;
|
||||
}
|
||||
|
||||
export interface ESError {
|
||||
root_cause?: ESErrorCausedBy[];
|
||||
caused_by?: ESErrorCausedBy;
|
||||
}
|
||||
|
||||
function extractCausedByChain(
|
||||
causedBy: ESErrorCausedBy = {},
|
||||
accumulator: string[] = []
|
||||
): string[] {
|
||||
const { reason, caused_by: innerCausedBy } = causedBy;
|
||||
|
||||
if (reason) {
|
||||
accumulator.push(reason);
|
||||
}
|
||||
|
||||
if (innerCausedBy) {
|
||||
return extractCausedByChain(innerCausedBy, accumulator);
|
||||
}
|
||||
|
||||
return accumulator;
|
||||
}
|
||||
|
||||
/**
|
||||
* Identified causes for ES Error
|
||||
*
|
||||
* @param err Object Error thrown by ES JS client
|
||||
* @return ES error cause
|
||||
*/
|
||||
export function identifyEsError(err: { response: string }) {
|
||||
const { response } = err;
|
||||
|
||||
if (response) {
|
||||
const { error } = JSON.parse(response) as { error?: ESError };
|
||||
if (error) {
|
||||
const { root_cause: rootCause = [], caused_by: causedBy } = error;
|
||||
|
||||
return [
|
||||
...extractCausedByChain(causedBy),
|
||||
...rootCause.reduce(
|
||||
(acc: string[], innerRootCause) => extractCausedByChain(innerRootCause, acc),
|
||||
[]
|
||||
),
|
||||
];
|
||||
}
|
||||
}
|
||||
return [];
|
||||
}
|
|
@ -188,5 +188,34 @@ describe('TaskManager', () => {
|
|||
|
||||
expect(claim).not.toHaveBeenCalled();
|
||||
});
|
||||
|
||||
/**
|
||||
* This handles the case in which Elasticsearch has had inline script disabled.
|
||||
* This is achieved by setting the `script.allowed_types` flag on Elasticsearch to `none`
|
||||
*/
|
||||
test('handles failure due to inline scripts being disabled', () => {
|
||||
const logger = mockLogger();
|
||||
const claim = jest.fn(() => {
|
||||
throw Object.assign(new Error(), {
|
||||
msg: '[illegal_argument_exception] cannot execute [inline] scripts',
|
||||
path: '/.kibana_task_manager/_update_by_query',
|
||||
query: {
|
||||
ignore_unavailable: true,
|
||||
refresh: true,
|
||||
max_docs: 200,
|
||||
conflicts: 'proceed',
|
||||
},
|
||||
body:
|
||||
'{"query":{"bool":{"must":[{"term":{"type":"task"}},{"bool":{"must":[{"bool":{"should":[{"bool":{"must":[{"term":{"task.status":"idle"}},{"range":{"task.runAt":{"lte":"now"}}}]}},{"bool":{"must":[{"bool":{"should":[{"term":{"task.status":"running"}},{"term":{"task.status":"claiming"}}]}},{"range":{"task.retryAt":{"lte":"now"}}}]}}]}},{"bool":{"should":[{"exists":{"field":"task.interval"}},{"bool":{"must":[{"term":{"task.taskType":"vis_telemetry"}},{"range":{"task.attempts":{"lt":3}}}]}},{"bool":{"must":[{"term":{"task.taskType":"lens_telemetry"}},{"range":{"task.attempts":{"lt":3}}}]}},{"bool":{"must":[{"term":{"task.taskType":"actions:.server-log"}},{"range":{"task.attempts":{"lt":1}}}]}},{"bool":{"must":[{"term":{"task.taskType":"actions:.slack"}},{"range":{"task.attempts":{"lt":1}}}]}},{"bool":{"must":[{"term":{"task.taskType":"actions:.email"}},{"range":{"task.attempts":{"lt":1}}}]}},{"bool":{"must":[{"term":{"task.taskType":"actions:.index"}},{"range":{"task.attempts":{"lt":1}}}]}},{"bool":{"must":[{"term":{"task.taskType":"actions:.pagerduty"}},{"range":{"task.attempts":{"lt":1}}}]}},{"bool":{"must":[{"term":{"task.taskType":"actions:.webhook"}},{"range":{"task.attempts":{"lt":1}}}]}}]}}]}}]}},"sort":{"_script":{"type":"number","order":"asc","script":{"lang":"expression","source":"doc[\'task.retryAt\'].value || doc[\'task.runAt\'].value"}}},"seq_no_primary_term":true,"script":{"source":"ctx._source.task.ownerId=params.ownerId; ctx._source.task.status=params.status; ctx._source.task.retryAt=params.retryAt;","lang":"painless","params":{"ownerId":"kibana:5b2de169-2785-441b-ae8c-186a1936b17d","retryAt":"2019-10-31T13:35:43.579Z","status":"claiming"}}}',
|
||||
statusCode: 400,
|
||||
response:
|
||||
'{"error":{"root_cause":[{"type":"illegal_argument_exception","reason":"cannot execute [inline] scripts"}],"type":"search_phase_execution_exception","reason":"all shards failed","phase":"query","grouped":true,"failed_shards":[{"shard":0,"index":".kibana_task_manager_1","node":"24A4QbjHSK6prvtopAKLKw","reason":{"type":"illegal_argument_exception","reason":"cannot execute [inline] scripts"}}],"caused_by":{"type":"illegal_argument_exception","reason":"cannot execute [inline] scripts","caused_by":{"type":"illegal_argument_exception","reason":"cannot execute [inline] scripts"}}},"status":400}',
|
||||
});
|
||||
});
|
||||
|
||||
claimAvailableTasks(claim, 10, logger);
|
||||
|
||||
sinon.assert.calledWithMatch(logger.warn, /inline scripts/);
|
||||
});
|
||||
});
|
||||
});
|
||||
|
|
|
@ -28,6 +28,7 @@ import {
|
|||
OwnershipClaimingOpts,
|
||||
ClaimOwnershipResult,
|
||||
} from './task_store';
|
||||
import { identifyEsError } from './lib/identify_es_error';
|
||||
|
||||
export interface TaskManagerOpts {
|
||||
logger: Logger;
|
||||
|
@ -256,20 +257,31 @@ export async function claimAvailableTasks(
|
|||
logger: Logger
|
||||
) {
|
||||
if (availableWorkers > 0) {
|
||||
const { docs, claimedTasks } = await claim({
|
||||
size: availableWorkers,
|
||||
claimOwnershipUntil: intervalFromNow('30s')!,
|
||||
});
|
||||
try {
|
||||
const { docs, claimedTasks } = await claim({
|
||||
size: availableWorkers,
|
||||
claimOwnershipUntil: intervalFromNow('30s')!,
|
||||
});
|
||||
|
||||
if (docs.length !== claimedTasks) {
|
||||
logger.warn(
|
||||
`[Task Ownership error]: (${claimedTasks}) tasks were claimed by Kibana, but (${docs.length}) tasks were fetched`
|
||||
);
|
||||
if (docs.length !== claimedTasks) {
|
||||
logger.warn(
|
||||
`[Task Ownership error]: (${claimedTasks}) tasks were claimed by Kibana, but (${docs.length}) tasks were fetched`
|
||||
);
|
||||
}
|
||||
return docs;
|
||||
} catch (ex) {
|
||||
if (identifyEsError(ex).includes('cannot execute [inline] scripts')) {
|
||||
logger.warn(
|
||||
`Task Manager cannot operate when inline scripts are disabled in Elasticsearch`
|
||||
);
|
||||
} else {
|
||||
throw ex;
|
||||
}
|
||||
}
|
||||
return docs;
|
||||
} else {
|
||||
logger.info(
|
||||
`[Task Ownership]: Task Manager has skipped Claiming Ownership of available tasks at it has ran out Available Workers. If this happens often, consider adjusting the "xpack.task_manager.max_workers" configuration.`
|
||||
);
|
||||
}
|
||||
logger.info(
|
||||
`[Task Ownership]: Task Manager has skipped Claiming Ownership of available tasks at it has ran out Available Workers. If this happens often, consider adjusting the "xpack.task_manager.max_workers" configuration.`
|
||||
);
|
||||
return [];
|
||||
}
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue