Introduction
Historically opReports has relied solely on being run on demand via the GUI or command line (or periodically with cron). In version 3.0 we've added another feature that extends these capabilities: the opReports Scheduler. The scheduler is meant to act a bit like a combination of "cron" and "at" (a venerable but not especially well known Unix tool, companion of cron) for opReports.
The basic idea is that you control the scheduler by defining opReports "jobs" to be taken care of (once, at a particular time in the future, or recurring with specific frequencies and start/end times/dates).
How to define Report Schedules
Each scheduled job is defined by a separate JSON file. The scheduler expects these files in /usr/local/omk/conf/schedule (but this is configurable - see <omk_schedule> in conf/opCommon.nmis). As of version 3.0 schedule files must be managed using an external editor, but support for editing schedules from the GUI is already planned.
All job files must be called <something>.json, and only letters, digits, dot, underscore and hyphen are allowed in the file name. All other files are silently ignored. You can thus disable a job by naming it something like !disabled_job.json. opReports ships with a number of example schedule files (all set to be inactive) and a brief README file in the conf/schedule directory.
A job file must consist of one "hash" datastructure, with the following properties:
| Property | Example | Description |
|---|---|---|
type | "util" "wan" | The report type to generate. Required. See the opReports Report Descriptions for a list of available types. |
| active | "true" or "false" | Whether this schedule is active or not. Only active schedules can create new reports, inactive schedules are consulted only for pruning of old report data. If active is not specified, then the schedule is inactive. |
| description | "some text" | A free-form description of the scheduled job. Optional. |
| frequency | "daily", "weekly", "monthly" or "yearly" | Not present for one-off reports. |
| start | "1-14-2015 20:00:00", "mon 10:30", various others depending on the frequency | The date + time the report collection is to start at. Required. |
| end | The date + time the report collection is to end at. Same format as start. | |
| from | 13 or any integer between 0 and 24 | Start of Business Hours reporting. Optional. Default: 0. See the Formats for Report Periods and Frequencies page for details. |
| to | End of Business Hours reporting. Optional. Default: 24. | |
| exclude | "fri-sun", "sat-mon" | Business week selection. Optional, supported in opReports 3.0.8 and newer. |
| node_intf_list | /some/where/some.file | The file describing the nodes and interfaces to collect. See section below for format. |
| util_threshold | 80 | The desired interface utilisation threshold value (in percent), default: 80 (percent). |
| util_threshold_mincount | 1 | The desired minimum number of threshold exceedences for flagging the interface as over-limit, default: 1. |
format
| a hash with the keys "html", "csv" and "xlsx", and values true/false or 1/0. e.g. { "html":0, "csv": true, "xlsx": false} | The formats the report should be saved in. |
| outputdir | /some/where/ | Where the report should be saved. This has to be a full path pointing to a directory. GUI-scheduled reports will always use one of the configured directories (see opreports_output_dirs in opCommon.nmis), and it is recommended that you use only these even if editing schedule files by hand.Every report schedule can use a different output directory. |
| naming | "simple" or "precise" | What naming scheme should be used for the output files. Default is "simple". Mainly important if you plan to manually work with saved report files. |
sourcesand (at most) one of node_list, node_regexp, node_group, node_intf_list or node_intf_type_list | "everything" | See How to select Nodes (and Interfaces) for details. |
| keep_for | 47 | How many days to keep an old report instance. If set to zero, the report is not expired. If not set, then the defaults for the report type are applied (configuration entry "default_report_keep_for"). |
| options | { "title" : "My Custom ReportTitle" }
| Optional settings. Some are specific to particular report types. See the section on Optional Settings below. |
| control_nmis | "true" or "false" | Whether NMIS should be "remote controlled" or left in peace. See section on Remote Controlling NMIS below. Only relevant for one-off reports. |
| nmis_options | "mthread=true maxthreads=15" | Options to be given to |
| target_audience_group | "HQ" | If present and a known NMIS group name, then the generated report can be viewed only by users who are members of this group (and the administrator). If not present then report viewing is not limited to particular groups. |
Besides these user-definable properties, the scheduler also manages certain others for internal use, and you should not modify these! (Currently these include the properties "uuid", "in_progress", "completed" and all properties named "actual_<something>".)
Scheduler Operation
When the scheduler starts it looks for jobs that are in need of actions (ie. the wanted reporting period has just passed and the relevant report hasn't been created yet) and handles them sequentially. When the scheduled job is finished, the job file is not removed and remains behind (regardless of whether this is a one-off or recurring schedule). After that all known saved reports are checked for pruning: if the report was created with a non-zero keep_for property, and is older than this cutoff then it is removed. This affects only report instances, not the job schedules.
The opReports scheduler is not a long-lived daemon process, but rather meant to be run suitably frequently from cron. The scheduler does log to opReports.log, but also prints error messages of higher urgency to STDERR, where cron will pick them up and mail them to root (or whatever recipient you have set up). By default opReports creates an /etc/cron.d/opreports which triggers the scheduler once every hour.
Remote-Controlling NMIS
This experimental feature is primarily meant for Interface Utilisation Reports at this time and works only for one-off jobs. If your job has control_nmis true AND a node_intf_list file, then the following actions are taken:
- When the job starts, NMIS is reconfigured with
global_collecttrue and with ONLY the wanted nodes and interfaces (identified by the node_intf_list file) set tocollecttrue. All other interface collection activity is disabled withglobal_nocollect_ifDescr. Annmis.pl type=updaterun will then be performed, and the NMIS data collection commences as normal. - When the job completes, NMIS is disabled, with
global_collectset to false. - This clearly interferes with recurring schedules and should therefore only be used in isolation.
As of version 3.0.2 some limitations apply for this feature:
- The NMIS remote control feature is not supported for recurring report schedules, only for one-off reports.
- The only possible extra action to take at the beginning of a job is enabling NMIS and setting the listed interfaces and nodes for collection.
Optional Settings
The options property can hold sub-properties that fine-tune opReports' behaviour. There are a few common options that are shared by all reports:
- Option title lets you specify a custom title for the report. The value must be a string.
- Option
homelinklets you specify a custom banner, to be displayed only for a standalone HTML report and above the actual report (title and body).
The value of this option must be a string that contains the raw HTML to be shown. - Option
emaillets you specify one or more recipients for report delivery by email. The value must be a comma separated list of email addresses. opReports will email the report to all listed email addresses, with all selected output formats as attachments. - Option
node_naminglets you change how node names are displayed.
If not set (the default), nodes are displayed with their plain node name.
If set todisplay_name, the NMIS propertydisplay_namewill be used if it's present for a node; if not, the plain node name is used.
This is supported in opReports 3.0.8 and NMIS 8.6 onwards.
The WAN report supports option wan_report_level, whose value selects which report detail level should be chosen. This can be given either as the index of the level or as its name. The page about WAN report detail levels describes this feature in detail.
The Node Health report has option exceptions, which when set to true, changes the report to only display health exceptions and suppress nodes with ok health values.
The Utilisation report recognizes the following four options:
| Option | Description |
|---|---|
| show_threshold | Default: true. If this is false, then no over threshold counts and exceptions are shown. Instead the interface bandwidths, traffic, utilisation and short report period are shown. |
| show_only_util | Default: false. Ignored if show_threshold isn't false. If this is set to true, then opReports omits the bandwidth and traffic columns and shows only the utilisation (and short period). |
| util_threshold | Default: 80 (%). Defines the level of utilisation above which it counts as exception. |
| util_threshold_mincount | Default: 1. How often the interface utilisation has to be above the threshold value for the interface to be flagged as in exceptional state. |