梳理需要监控的指标指标数据获取方案（获取方案，性能评估）开发方案（开发结构设计、sops/engine） <l

<div class="highlight highlight-source-shell notranslate position-relative overflow-auto" dir="auto"

执行失败任务sql <div class="highlight highlight-source-sql notranslate position-relati

系统任务执行状态自监控能力支持 about bk-sops HOT 2 OPEN

normal-wls commented on June 1, 2024

系统任务执行状态自监控能力支持

from bk-sops.

Comments (2)

lTimej commented on June 1, 2024

监控指标：
    1.执行中失败的任务
    2.执行中的任务

from bk-sops.

lTimej commented on June 1, 2024

执行失败任务sql

SELECT
	* 
FROM
	taskflow3_taskflowinstance AS t1
	INNER JOIN (
	SELECT
		pp.id 
	FROM
		pipeline_pipelineinstance AS pp
		INNER JOIN eri_state AS es ON pp.instance_id = es.root_id 
	WHERE
		es.NAME = "Failed" 
		AND pp.is_started = 1 
		AND pp.is_finished = 0 
		AND pp.is_revoked = 0 
		AND pp.is_expired = 0 
		AND is_deleted = 0 
	GROUP BY
		es.root_id,
		pp.id 
	ORDER BY
		pp.id 
	) AS t2 ON t1.pipeline_instance_id = t2.id;

执行失败任务ORM

# 过滤出失败的节点，根据节点分组
        states = [state["root_id"] for state in State.objects.filter(name="FAILED").annotate(count=Count("root_id")).values("root_id")]
        # 查出失败的节点任务
        tasks = TaskFlowInstance.objects.filter(
            pipeline_instance__instance_id__in=states,
            pipeline_instance__is_started=True,
            pipeline_instance__is_finished=False,
            pipeline_instance__is_revoked=False,
            pipeline_instance__is_expired=False,
            pipeline_instance__is_deleted=False,
            is_deleted=False,
        ).values(
            "project__name", 
            "pipeline_instance__name",
            "pipeline_instance__start_time",
        )

执行中任务sql

SELECT
	* 
FROM
	taskflow3_taskflowinstance AS t1
	INNER JOIN (
	SELECT
		pp.id,
		COUNT( es.root_id ) AS total_count,
		SUM( CASE WHEN es.NAME = 'Failed' THEN 1 ELSE 0 END ) AS failed_count,
		SUM( CASE WHEN es.NAME = 'FINISHED' THEN 1 ELSE 0 END ) AS finished_count 
	FROM
		pipeline_pipelineinstance AS pp
		INNER JOIN eri_state AS es ON pp.instance_id = es.root_id 
	WHERE
		pp.is_started = 1 
		AND pp.is_finished = 0 
		AND pp.is_revoked = 0 
		AND pp.is_expired = 0 
		AND is_deleted = 0 
	GROUP BY
		es.root_id,
		pp.id 
	HAVING
		failed_count = 0 
		AND finished_count < total_count 
	ORDER BY
		pp.id 
	) AS t2 ON t1.pipeline_instance_id = t2.id;

执行中任务ORM

states = [
            state["root_id"] 
            for state in (
                State.objects.values("root_id").
                annotate(total_count=Count("root_id")).
                annotate(failed_count=Count("root_id", filter=Q(name='FAILED'))).
                annotate(finished_count=Count("root_id",filter=Q(name='FINISHED')))
            )
            if state["failed_count"] == 0 and state["finished_count"] < state["total_count"]
        ]
        tasks = TaskFlowInstance.objects.filter(
            pipeline_instance__instance_id__in=states,
            pipeline_instance__is_started=True,
            pipeline_instance__is_finished=False,
            pipeline_instance__is_revoked=False,
            pipeline_instance__is_expired=False,
            pipeline_instance__is_deleted=False,
            is_deleted=False,
        ).values(
            "project__name", 
            "pipeline_instance__name",
            "pipeline_instance__start_time",
        )