com.atlassian.scheduler.caesium.spi

Interface CaesiumSchedulerConfiguration

All Superinterfaces:

com.atlassian.scheduler.core.spi.SchedulerServiceConfiguration

public interface CaesiumSchedulerConfiguration
extends com.atlassian.scheduler.core.spi.SchedulerServiceConfiguration

Provides custom configuration settings for the Caesium scheduler service.

Since:

v0.0.1

Method Summary

Modifier and Type	Method and Description
int	refreshClusteredJobsIntervalInMinutes() Returns the frequency with which the scheduler will perform routine cluster maintenance.
boolean	useFineGrainedSchedules() If true, then the scheduler will accept schedules that are more fine-grained than once per minute.
boolean	useQuartzJobDataMapMigration() If true, then the serialized parameter maps returned by ClusteredJobDao will be allowed to contain instances of Quartz's JobDataMap class, which will be removed transparently.
int	workerThreadCount() Returns the number of worker threads that will accept jobs from the queue.

Methods inherited from interface com.atlassian.scheduler.core.spi.SchedulerServiceConfiguration

getDefaultTimeZone

Method Detail

refreshClusteredJobsIntervalInMinutes

int refreshClusteredJobsIntervalInMinutes()

Returns the frequency with which the scheduler will perform routine cluster maintenance.

If the value is positive, then the scheduler will implicitly register a local maintenance task that calls CaesiumSchedulerService.refreshClusteredJobs() at the specified interval.

If the value is not positive, then the application must make its own arrangements to notify other cluster nodes about new clustered jobs. For example, it might use a message passing service to explicitly inform other nodes that a new cluster-wide job has been scheduled. The nodes that receive such a message could then call CaesiumSchedulerService.refreshClusteredJob(JobId) to ensure that the new job is included in the scheduler's work queue.

It would also be appropriate to disable this for any single-node application, as the database poll to check for updated clustered jobs would just be wasteful.

Returns:

the frequency (in minutes) at which the scheduler will automatically poll for changes to clustered jobs, or a non-positive value to disable this feature

See Also:

CaesiumSchedulerService.refreshClusteredJobs(), CaesiumSchedulerService.refreshClusteredJob(JobId)

workerThreadCount

int workerThreadCount()

Returns the number of worker threads that will accept jobs from the queue.

If this value is not positive, then the scheduler will choose the number of threads that it will use. The algorithm used to select the default thread count is left unspecified, so it is recommended that applications provide an explicit value rather than accept the default.

Returns:

the number of worker threads to use, or a non-positive value to accept the default value

useQuartzJobDataMapMigration

boolean useQuartzJobDataMapMigration()

If true, then the serialized parameter maps returned by ClusteredJobDao will be allowed to contain instances of Quartz's JobDataMap class, which will be removed transparently.

This is intended to make it easier to migrate from Quartz to Caesium by allowing the job data to be cleaned lazily instead of requiring the Quartz classes to be removed from the data up front. There may be a minor performance penalty to this support, as the deserialization code must intercept the requests for the Quartz classes by comparing the names of each class as it is resolved, but this is not expected to be significant.

Returns:

true to enable transparent migration of Quartz data returned by the ClusteredJobDao; false to disable this feature.

useFineGrainedSchedules

boolean useFineGrainedSchedules()

If true, then the scheduler will accept schedules that are more fine-grained than once per minute.

The atlassian-scheduler library is permitted to "adjust" schedules that would run more frequently than once a minute to prevent abuse. Setting this to true will disable this restriction, allowing intervals shorter than one minute and cron expressions that specify values other than a single integer from 0 to 59 for the seconds field.

Returns:

true to allow all valid schedule requests, regardless of how frequently they will run; false if schedules should be adjusted to prevent schedules from running more frequently than once a minute.