1 =======================
2 Ganeti monitoring agent
3 =======================
5 .. contents:: :depth: 4
7 This is a design document detailing the implementation of a Ganeti
8 monitoring agent report system, that can be queried by a monitoring
9 system to calculate health information for a Ganeti cluster.
11 Current state and shortcomings
12 ==============================
14 There is currently no monitoring support in Ganeti. While we don't want
15 to build something like Nagios or Pacemaker as part of Ganeti, it would
16 be useful if such tools could easily extract information from a Ganeti
17 machine in order to take actions (example actions include logging an
18 outage for future reporting or alerting a person or system about it).
23 Each Ganeti node should export a status page that can be queried by a
24 monitoring system. Such status page will be exported on a network port
25 and will be encoded in JSON (simple text) over HTTP.
27 The choice of JSON is obvious as we already depend on it in Ganeti and
28 thus we don't need to add extra libraries to use it, as opposed to what
29 would happen for XML or some other markup format.
31 Location of agent report
32 ------------------------
34 The report will be available from all nodes, and be concerned for all
35 node-local resources. This allows more real-time information to be
36 available, at the cost of querying all nodes.
41 The monitoring agent system will report on the following basic information:
44 - Instance disk status
45 - Status of storage for instances
46 - Ganeti daemons status, CPU usage, memory footprint
47 - Hypervisor resources report (memory, CPU, network interfaces)
48 - Node OS resources report (memory, CPU, network interfaces)
49 - Information from a plugin system
54 The report of the will be in JSON format, and it will present an array
56 Each report object will be produced by a specific data collector.
57 Each report object includes some mandatory fields, to be provided by all
61 The name of the data collector that produced this part of the report.
62 It is supposed to be unique inside a report.
65 The version of the data collector that produces this part of the
66 report. Built-in data collectors (as opposed to those implemented as
67 plugins) should have "B" as the version number.
70 The format of what is represented in the "data" field for each data
71 collector might change over time. Every time this happens, the
72 format_version should be changed, so that who reads the report knows
73 what format to expect, and how to correctly interpret it.
76 The time when the reported data were gathered. It has to be expressed
77 in nanoseconds since the unix epoch (0:00:00 January 01, 1970). If not
78 enough precision is available (or needed) it can be padded with
79 zeroes. If a report object needs multiple timestamps, it can add more
80 and/or override this one inside its own "data" section.
83 A collector can belong to a given category of collectors (e.g.: storage
84 collectors, daemon collector). This means that it will have to provide a
85 minumum set of prescribed fields, as documented for each category.
86 This field will contain the name of the category the collector belongs to,
87 if any, or just the ``null`` value.
90 Two kinds of collectors are possible:
91 `Performance reporting collectors`_ and `Status reporting collectors`_.
92 The respective paragraphs will describe them and the value of this field.
95 This field contains all the data generated by the specific data collector,
96 in its own independently defined format. The monitoring agent could check
97 this syntactically (according to the JSON specifications) but not
100 Here follows a minimal example of a report::
104 "name" : "TheCollectorIdentifier",
106 "format_version" : 1,
107 "timestamp" : 1351607182000000000,
110 "data" : { "plugin_specific_data" : "go_here" }
113 "name" : "AnotherDataCollector",
115 "format_version" : 7,
116 "timestamp" : 1351609526123854000,
117 "category" : "storage",
119 "data" : { "status" : { "code" : 1,
120 "message" : "Error on disk 2"
122 "plugin_specific" : "data",
123 "some_late_data" : { "timestamp" : 1351609526123942720,
130 Performance reporting collectors
131 ++++++++++++++++++++++++++++++++
133 These collectors only provide data about some component of the system, without
134 giving any interpretation over their meaning.
136 The value of the ``kind`` field of the report will be ``0``.
138 Status reporting collectors
139 +++++++++++++++++++++++++++
141 These collectors will provide information about the status of some
142 component of ganeti, or managed by ganeti.
144 The value of their ``kind`` field will be ``1``.
146 The rationale behind this kind of collectors is that there are some situations
147 where exporting data about the underlying subsystems would expose potential
148 issues. But if Ganeti itself is able (and going) to fix the problem, conflicts
149 might arise between Ganeti and something/somebody else trying to fix the same
151 Also, some external monitoring systems might not be aware of the internals of a
152 particular subsystem (e.g.: DRBD) and might only exploit the high level
153 response of its data collector, alerting an administrator if anything is wrong.
154 Still, completely hiding the underlying data is not a good idea, as they might
155 still be of use in some cases. So status reporting plugins will provide two
156 output modes: one just exporting a high level information about the status,
157 and one also exporting all the data they gathered.
158 The default output mode will be the status-only one. Through a command line
159 parameter (for stand-alone data collectors) or through the HTTP request to the
161 (when collectors are executed as part of it) the verbose output mode providing
162 all the data can be selected.
164 When exporting just the status each status reporting collector will provide,
165 in its ``data`` section, at least the following field:
168 summarizes the status of the component being monitored and consists of two
172 It assumes a numeric value, encoded in such a way to allow using a bitset
173 to easily distinguish which states are currently present in the whole cluster.
174 If the bitwise OR of all the ``status`` fields is 0, the cluster is
176 The status codes are as follows:
179 The collector can determine that everything is working as
183 Something is temporarily wrong but it is being automatically fixed by
185 There is no need of external intervention.
188 The collector has failed to understand whether the status is good or
189 bad. Further analysis is required. Interpret this status as a
190 potentially dangerous situation.
193 The collector can determine that something is wrong and Ganeti has no
194 way to fix it autonomously. External intervention is required.
197 A message to better explain the reason of the status.
198 The exact format of the message string is data collector dependent.
200 The field is mandatory, but the content can be an empty string if the
201 ``code`` is ``0`` (working as intended) or ``1`` (being fixed
204 If the status code is ``2``, the message should specify what has gone
206 If the status code is ``4``, the message shoud explain why it was not
207 possible to determine a proper status.
209 The ``data`` section will also contain all the fields describing the gathered
210 data, according to a collector-specific format.
215 At the moment each node knows which instances are running on it, which
216 instances it is primary for, but not the cause why an instance might not
217 be running. On the other hand we don't want to distribute full instance
218 "admin" status information to all nodes, because of the performance
219 impact this would have.
221 As such we propose that:
223 - Any operation that can affect instance status will have an optional
224 "reason" attached to it (at opcode level). This can be used for
225 example to distinguish an admin request, from a scheduled maintenance
226 or an automated tool's work. If this reason is not passed, Ganeti will
227 just use the information it has about the source of the request.
228 This reason information will be structured according to the
229 :doc:`Ganeti reason trail <design-reason-trail>` design document.
230 - RPCs that affect the instance status will be changed so that the
231 "reason" and the version of the config object they ran on is passed to
232 them. They will then export the new expected instance status, together
233 with the associated reason and object version to the status report
234 system, which then will export those themselves.
236 Monitoring and auditing systems can then use the reason to understand
237 the cause of an instance status, and they can use the timestamp to
238 understand the freshness of their data even in the absence of an atomic
239 cross-node reporting: for example if they see an instance "up" on a node
240 after seeing it running on a previous one, they can compare these values
241 to understand which data is freshest, and repoll the "older" node. Of
242 course if they keep seeing this status this represents an error (either
243 an instance continuously "flapping" between nodes, or an instance is
244 constantly up on more than one), which should be reported and acted
247 The instance status will be on each node, for the instances it is
248 primary for, and its ``data`` section of the report will contain a list
249 of instances, named ``instances``, with at least the following fields for
253 The name of the instance.
256 The UUID of the instance (stable on name change).
259 The status of the instance (up/down/offline) as requested by the admin.
262 The actual status of the instance. It can be ``up``, ``down``, or
263 ``hung`` if the instance is up but it appears to be completely stuck.
266 The uptime of the instance (if it is up, "null" otherwise).
269 The timestamp of the last known change to the instance state.
272 The last known reason for state change of the instance, described according
273 to the JSON representation of a reason trail, as detailed in the :doc:`reason
274 trail design document <design-reason-trail>`.
277 It represents the status of the instance, and its format is the same as that
278 of the ``status`` field of `Status reporting collectors`_.
280 Each hypervisor should provide its own instance status data collector, possibly
281 with the addition of more, specific, fields.
282 The ``category`` field of all of them will be ``instance``.
283 The ``kind`` field will be ``1``.
285 Note that as soon as a node knows it's not the primary anymore for an
286 instance it will stop reporting status for it: this means the instance
287 will either disappear, if it has been deleted, or appear on another
288 node, if it's been moved.
290 The ``code`` of the ``status`` field of the report of the Instance status data
294 if ``status`` is ``0`` for all the instances it is reporting about.
302 The storage collectors will be a series of data collectors
303 that will gather data about storage for the current node. The collection
304 will be performed at different granularity and abstraction levels, from
305 the physical disks, to partitions, logical volumes and to the specific
306 storage types used by Ganeti itself (drbd, rbd, plain, file).
308 The ``name`` of each of these collector will reflect what storage type each of
311 The ``category`` field of these collector will be ``storage``.
313 The ``kind`` field will depend on the specific collector.
315 Each ``storage`` collector's ``data`` section will provide collector-specific
318 In case of error, the ``message`` subfield of the ``status`` field of the
319 report of the instance status collector will disclose the nature of the error
320 as a type specific information. Examples of these are "backend pv unavailable"
321 for lvm storage, "unreachable" for network based storage or "filesystem error"
322 for filesystem based implementations.
327 This storage data collector will gather information about the status of the
328 disks installed in the system, as listed in the /proc/diskstats file. This means
329 that not only physical hard drives, but also ramdisks and loopback devices will
332 Its ``kind`` in the report will be ``0`` (`Performance reporting collectors`_).
334 Its ``category`` field in the report will contain the value ``storage``.
336 When executed in verbose mode, the ``data`` section of the report of this
337 collector will be a list of items, each representing one disk, each providing
338 the following fields:
341 The major number of the device.
344 The minor number of the device.
347 The name of the device.
350 This is the total number of reads completed successfully.
353 Reads which are adjacent to each other may be merged for efficiency. Thus
354 two 4K reads may become one 8K read before it is ultimately handed to the
355 disk, and so it will be counted (and queued) as only one I/O. This field
356 specifies how often this was done.
359 This is the total number of sectors read successfully.
362 This is the total number of milliseconds spent by all reads.
365 This is the total number of writes completed successfully.
368 Writes which are adjacent to each other may be merged for efficiency. Thus
369 two 4K writes may become one 8K read before it is ultimately handed to the
370 disk, and so it will be counted (and queued) as only one I/O. This field
371 specifies how often this was done.
374 This is the total number of sectors written successfully.
377 This is the total number of milliseconds spent by all writes
380 The number of I/Os currently in progress.
381 The only field that should go to zero, it is incremented as requests are
382 given to appropriate struct request_queue and decremented as they finish.
385 The number of milliseconds spent doing I/Os. This field increases so long
386 as field ``IOs`` is nonzero.
389 The weighted number of milliseconds spent doing I/Os.
390 This field is incremented at each I/O start, I/O completion, I/O merge,
391 or read of these stats by the number of I/Os in progress (field ``IOs``)
392 times the number of milliseconds spent doing I/O since the last update of
393 this field. This can provide an easy measure of both I/O completion time
394 and the backlog that may be accumulating.
399 This data collector will run only on nodes where DRBD is actually
400 present and it will gather information about DRBD devices.
402 Its ``kind`` in the report will be ``1`` (`Status reporting collectors`_).
404 Its ``category`` field in the report will contain the value ``storage``.
406 When executed in verbose mode, the ``data`` section of the report of this
407 collector will provide the following fields:
410 Information about the DRBD version number, given by a combination of
411 any (but at least one) of the following fields:
414 The DRBD driver version.
417 The API version number.
420 The protocol version.
423 The version of the source files.
426 Git hash of the source files.
429 Who built the binary, and, optionally, when.
432 A list of structures, each describing a DRBD device (a minor) and containing
433 the following fields:
436 The device minor number.
439 The state of the connection. If it is "Unconfigured", all the following
440 fields are not present.
443 The role of the local resource.
446 The role of the remote resource.
449 The status of the local disk.
452 The status of the remote disk.
454 ``replicationProtocol``
455 The replication protocol being used.
458 The input/output flags.
461 The performance indicators. This field will contain the following
465 KiB of data sent on the network.
468 KiB of data received from the network.
471 KiB of data written on local disk.
474 KiB of date read from the local disk.
477 Number of updates of the activity log.
480 Number of updates to the bitmap area of the metadata.
483 Number of open requests to the local I/O subsystem.
486 Number of requests sent to the partner but not yet answered.
489 Number of requests received by the partner but still to be answered.
491 ``applicationPending``
492 Num of block input/output requests forwarded to DRBD but that have not yet
496 (Optional) Number of epoch objects. Not provided by all DRBD versions.
499 (Optional) Currently used write ordering method. Not provided by all DRBD
503 (Optional) KiB of storage currently out of sync. Not provided by all DRBD
507 (Optional) The status of the synchronization of the disk. This is present
508 only if the disk is being synchronized, and includes the following fields:
511 The percentage of synchronized data.
514 How far the synchronization is. Written as "x/y", where x and y are
515 integer numbers expressed in the measurement unit stated in
519 The measurement unit for the progress indicator.
522 The expected time before finishing the synchronization.
525 The speed of the synchronization.
528 The desiderd speed of the synchronization.
531 The measurement unit of the ``speed`` and ``want`` values. Expressed
535 The name of the Ganeti instance this disk is associated to.
538 Ganeti daemons status
539 +++++++++++++++++++++
541 Ganeti will report what information it has about its own daemons.
542 This should allow identifying possible problems with the Ganeti system itself:
543 for example memory leaks, crashes and high resource utilization should be
544 evident by analyzing this information.
546 The ``kind`` field will be ``1`` (`Status reporting collectors`_).
548 Each daemon will have its own data collector, and each of them will have
549 a ``category`` field valued ``daemon``.
551 When executed in verbose mode, their data section will include at least:
554 The amount of used memory.
557 The measurement unit used for the memory.
560 The uptime of the daemon.
563 How much cpu the daemon is using (percentage).
565 Any other daemon-specific information can be included as well in the ``data``
568 Hypervisor resources report
569 +++++++++++++++++++++++++++
571 Each hypervisor has a view of system resources that sometimes is
572 different than the one the OS sees (for example in Xen the Node OS,
573 running as Dom0, has access to only part of those resources). In this
574 section we'll report all information we can in a "non hypervisor
575 specific" way. Each hypervisor can then add extra specific information
576 that is not generic enough be abstracted.
578 The ``kind`` field will be ``0`` (`Performance reporting collectors`_).
580 Each of the hypervisor data collectory will be of ``category``: ``hypervisor``.
582 Node OS resources report
583 ++++++++++++++++++++++++
585 Since Ganeti assumes it's running on Linux, it's useful to export some
586 basic information as seen by the host system.
588 The ``category`` field of the report will be ``null``.
590 The ``kind`` field will be ``0`` (`Performance reporting collectors`_).
592 The ``data`` section will include:
595 The number of available cpus.
598 A list with one element per cpu, showing its average load.
601 The current view of memory (free, used, cached, etc.)
604 A list with one element per filesystem, showing a summary of the
605 total/available space.
608 A list with one element per network interface, showing the amount of
609 sent/received data, error rate, IP address of the interface, etc.
612 A map using the name of a component Ganeti interacts (Linux, drbd,
613 hypervisor, etc) as the key and its version number as the value.
615 Note that we won't go into any hardware specific details (e.g. querying a
616 node RAID is outside the scope of this, and can be implemented as a
617 plugin) but we can easily just report the information above, since it's
618 standard enough across all systems.
623 .. include:: monitoring-query-format.rst
625 Instance disk status propagation
626 --------------------------------
628 As for the instance status Ganeti has now only partial information about
629 its instance disks: in particular each node is unaware of the disk to
630 instance mapping, that exists only on the master.
632 For this design doc we plan to fix this by changing all RPCs that create
633 a backend storage or that put an already existing one in use and passing
634 the relevant instance to the node. The node can then export these to the
635 status reporting tool.
637 While we haven't implemented these RPC changes yet, we'll use Confd to
638 fetch this information in the data collectors.
643 The monitoring system will be equipped with a plugin system that can
644 export specific local information through it.
646 The plugin system is expected to be used by local installations to
647 export any installation specific information that they want to be
648 monitored, about either hardware or software on their systems.
650 The plugin system will be in the form of either scripts or binaries whose output
651 will be inserted in the report.
653 Eventually support for other kinds of plugins might be added as well, such as
654 plain text files which will be inserted into the report, or local unix or
655 network sockets from which the information has to be read. This should allow
656 most flexibility for implementing an efficient system, while being able to keep
657 it as simple as possible.
662 In order to ease testing as well as to make it simple to reuse this
663 subsystem it will be possible to run just the "data collectors" on each
664 node without passing through the agent daemon.
666 If a data collector is run independently, it should print on stdout its
667 report, according to the format corresponding to a single data collector
668 report object, as described in the previous paragraphs.
673 In order to be able to report information fast the monitoring agent
674 daemon will keep an in-memory or on-disk cache of the status, which will
675 be returned when queries are made. The status system will then
676 periodically check resources to make sure the status is up to date.
678 Different parts of the report will be queried at different speeds. These
680 - how often they vary (or we expect them to vary)
681 - how fast they are to query
682 - how important their freshness is
684 Of course the last parameter is installation specific, and while we'll
685 try to have defaults, it will be configurable. The first two instead we
686 can use adaptively to query a certain resource faster or slower
687 depending on those two parameters.
689 When run as stand-alone binaries, the data collector will not using any
690 caching system, and just fetch and return the data immediately.
695 The status daemon will be implemented as a standalone Haskell daemon. In
696 the future it should be easy to merge multiple daemons into one with
697 multiple entry points, should we find out it saves resources and doesn't
698 impact functionality.
700 The libekg library should be looked at for easily providing metrics in
706 We will implement the agent system in this order:
708 - initial example data collectors (eg. for drbd and instance status).
709 - initial daemon for exporting data, integrating the existing collectors
711 - RPC updates for instance status reasons and disk to instance mapping
712 - cache layer for the daemon
713 - more data collectors
719 As a future step it can be useful to "centralize" all this reporting
720 data on a single place. This for example can be just the master node, or
721 all the master candidates. We will evaluate doing this after the first
722 node-local version has been developed and tested.
724 Another possible change is replacing the "read-only" RPCs with queries
725 to the agent system, thus having only one way of collecting information
726 from the nodes from a monitoring system and for Ganeti itself.
728 One extra feature we may need is a way to query for only sub-parts of
729 the report (eg. instances status only). This can be done by passing
730 arguments to the HTTP GET, which will be defined when we get to this
733 Finally the :doc:`autorepair system design <design-autorepair>`. system
734 (see its design) can be expanded to use the monitoring agent system as a
735 source of information to decide which repairs it can perform.
737 .. vim: set textwidth=72 :