接下来的几篇文章将列举和描述Cloud Foundry系统组件的相关指标。这些指标可以从Loggregator Firehose获得。
本文将以表格形式列举并描述Cloud Controller的指标。
Cloud Controller（CC）
指标名 | 详细描述 | diego_sync.invalid_desired_lrps Number of invalid DesiredLRPs found during CF apps and Diego DesiredLRPs periodic synchronization. Emitted every 30 seconds. failed_job_count.- Number of failed jobs in the - queue. This is the number of delayed jobs where the failed at column is populated with the time of the most recently failed attempt at the job. The failed job count is not specific to the jobs run by the Cloud Controller worker. By default, Cloud Controller deletes failed jobs after 31 days. Emitted every 30 seconds per VM. diego_sync.duration Time in milliseconds that it took to synchronize CF apps and Diego DesiredLRPs. Emitted every 30 seconds. failed_job_count.cc-generic Number of failed jobs in the cc-generic queue. By default, Cloud Controller deletes failed jobs after 31 days. Emitted every 30 seconds per VM. failed_job_count.total Number of failed jobs in all queues. By default, Cloud Controller deletes failed jobs after 31 days. Emitted every 30 seconds per VM. http_status.1XX Number of HTTP response status codes of type 1xx (informational). This resets when the Cloud Controller process is restarted and is incremented at the end of each request cycle. http_status.2XX Number of HTTP response status codes of type 2xx (success). This resets when the Cloud Controller process is restarted and is incremented at the end of each request cycle. Emitted for each Cloud Controller request. http_status.3XX Number of HTTP response status codes of type 3xx (redirection). This resets when the Cloud Controller process is restarted and is incremented at the end of each request cycle. Emitted for each Cloud Controller request. http_status.4XX Number of HTTP response status codes of type 4xx (client error). This resets when the Cloud Controller process is restarted and is incremented at the end of each request cycle. Emitted for each Cloud Controller request. http_status.5XX Number of HTTP response status codes of type 5xx (server error). This resets when the Cloud Controller process is restarted and is incremented at the end of each request cycle. job_queue_length.cc– Number of background jobs in the - queue that have yet to run for the first time. Emitted every 30 seconds per VM. job_queue_length.cc-generic Number of background jobs in the cc-generic queue that have yet to run for the first time. Emitted every 30 seconds per VM. job_queue_length.total Total number of background jobs in the queues that have yet to run for the first time. Emitted every 30 seconds per VM. log_count.all Total number of log messages, sum of messages of all severity levels. The count resets when the Cloud Controller process is restarted. Emitted every 30 seconds per VM. log_count.debug Number of log messages of severity “debug.” The count resets when the Cloud Controller process is restarted. Emitted every 30 seconds per VM. log_count.debug1 Not used. log_count.debug2 Number of log messages of severity “debug2.” The count resets when the Cloud Controller process is restarted. Emitted every 30 seconds per VM. log_count.error Number of log messages of severity “error.” Error is the most severe level. It is used for failures and during error handling. Most errors can be found under this log level, eg. failed unbinding a service, failed to cancel a task, Diego app crashed error, staging completion errors, staging errors, and resource not found. The count resets when the Cloud Controller process is restarted. Emitted every 30 seconds per VM. log_count.fatal Number of log messages of severity “fatal.” The count resets when the Cloud Controller process is restarted. Emitted every 30 seconds per VM. log_count.info Number of log messages of severity “info.” Examples of info messages are droplet created, copying package, uploading package, access denied due to insufficient scope, job logging, blobstore actions, staging requests, and app running requests. The count resets when the Cloud Controller process is restarted. Emitted every 30 seconds per VM. log_count.off Number of log messages of severity “off.” The count resets when the Cloud Controller process is restarted. Emitted every 30 seconds per VM. log_count.warn Number of log messages of severity “warn.” Warn is also used for failures and during error handling, eg. diagnostics written to file, failed to capture diagnostics, app rollback failed, service broker already deleted, and UAA token problems. The count resets when the Cloud Controller process is restarted. Emitted every 30 seconds per VM. requests.completed Number of requests that have been processed. Emitted for each Cloud Controller request. requests.outstanding Number of request that are currently being processed. Emitted for each Cloud Controller request. staging.requested Cumulative number of requests to start a staging task handled by each Cloud Controller. staging.succeeded Cumulative number of successful staging tasks handled by each Cloud Controller. Emitted every time a staging task completes successfully. staging.succeeded_duration Time in milliseconds that the successful staging task took to run. Emitted each time a staging task completes successfully. staging.failed Cumulative number of failed staging tasks handled by each Cloud Controller. Emitted every time a staging task fails. staging.failed_duration Time in milliseconds that the failed staging task took to run. Emitted each time a staging task fails. tasks_running.count Number of currently running tasks. Emitted every 30 seconds per VM. This metric is only seen in version 3 of the Cloud Foundry API. tasks_running.memory_in_mb Memory being consumed by all currently running tasks. Emitted every 30 seconds per VM. This metric is only seen in version 3 of the Cloud Foundry API. thread_info.event_machine.connection_count Number of open connections to event machine. Emitted every 30 seconds per VM. thread_info.event_machine.resultqueue.num_waiting Number of scheduled tasks in the result. Emitted every 30 seconds per VM. thread_info.event_machine.resultqueue.size Number of unscheduled tasks in the result. Emitted every 30 seconds per VM. thread_info.event_machine.threadqueue.num_waiting Number of scheduled tasks in the threadqueue. Emitted every 30 seconds per VM. thread_info.event_machine.threadqueue.size Number of unscheduled tasks in the threadqueue. Emitted every 30 seconds per VM. thread_info.thread_count Total number of threads that are either runnable or stopped. Emitted every 30 seconds per VM. total_users Total number of users ever created, including inactive users. Emitted every 10 minutes per VM. vcap_sinatra.recent_errors 50 most recent errors. DEPRECATED vitals.cpu Percentage of CPU used by the Cloud Controller process. Emitted every 30 seconds per VM. vitals.cpu_load_avg System CPU load averaged over the last 1 minute according to the OS. Emitted every 30 seconds per VM. vitals.mem_bytes The RSS bytes (resident set size) or real memory of the Cloud Controller process. Emitted every 30 seconds per VM. vitals.mem_free_bytes Total memory available according to the OS. Emitted every 30 seconds per VM. vitals.mem_used_bytes Total memory used (active + wired) according to the OS. Emitted every 30 seconds per VM. vitals.num_cores The number of CPUs of a host machine. Emitted every 30 seconds per VM. vitals.uptime The uptime of the Cloud Controller process in seconds. Emitted every 30 seconds per VM.
