Open-metrics endpoint for ONTAP and StorageGRID
Nightly builds may include bugs and other issues. You might want to use the stable releases instead.
:pushpin: This release is the same as 24.05.0 with a fix that makes the NFS Troubleshooting dashboards load in NAbox. If you are not using NAbox or you do not use the NFS trouble shooting dashboards, you can ignore this release.
Full Changelog: https://github.com/NetApp/harvest/compare/v24.05.0...v24.05.1
:pushpin: Highlights of this major release include:
Harvest supports consistency groups (CG) in the SnapMirror dashboard. Thanks to @Nikhita-13 for reporting this.
We've fixed an intermittent latency/ops spike problem caused by Harvest incorrectly handling ONTAP partial aggregation. This impacted all perf objects. A big thank you to @summertony15 for reporting this critical issue.
Harvest dashboards are compatible with Grafana 10.x.x versions.
:gem: LUN, Flexgroup and cDot dashboard updated to work with FSx. Some panels are blank because FSx does not have that data.
The credentials script supports providing both username and password. Thanks to @kbhalaki for reporting.
Harvest configuration file supports reading parameters from environment variables. Kudos to @wally007 for the suggestion.
Harvest includes remediation steps for EMS alerts.
:gem: New Dashboards:
NFS Troubleshooting which provides links to detailed dashboards. Thanks to RustyBrown for contributing these.Volume by SVM and Volume Deep Dive.:rocket: Performance Improvements:
:star: Several of the existing dashboards include new panels in this release:
:ear_of_rice: Harvest includes a new template to collect lock counts at the node, SVM, LIF, and volume levels.. Thanks to @troysmullerna for reporting.
:closed_book: Documentation Additions:
confpath to extend templates.Harvest supports embedded exporters in Harvest configuration. This means you can define your exporters in one place instead of multiple. Thanks to @wagneradrian92 for reporting.
Harvest supports exporting to multiple InfluxDB instances. Thanks to @figeac888 for reporting.
Node label metrics include HA partner details. Thanks to @johnwarlick for reporting.
:bangbang: IMPORTANT Release 24.05 removes duplicate quota metrics. If you wish to enable them, refer here.
:bulb: IMPORTANT After upgrading, don't forget to re-import your dashboards to get all the new enhancements and fixes. You can import them via the 'bin/harvest grafana import' CLI, from the Grafana UI, or from the 'Maintenance > Reset Harvest Dashboards' button in NAbox.
offbox_vscan_server and offbox_vscan until ONTAP 9.13.1. See ONTAP bug 1473892 for more details.IMPORTANT 7-mode filers that are not on the latest release of ONTAP may experience TLS connection issues with errors like tls: server selected unsupported protocol version 301. This is caused by a change in Go 1.18. The default for TLS client connections was changed to TLS 1.2 in Go 1.18. Please upgrade your 7-mode filers (recommended) or set tls_min_version: tls10 in your harvest.yml poller section. See #1007 for more details.
:metal: A big thanks to all the people who've opened issues, asked questions on Discord, and contributed code or dashboards this release:
@BrendonA667, @Nikhita-13, @WayneShen2, @derDaywalker, @faguayot, @figeac888, @jgasher, @johnwarlick, @kbhalaki, @rdecaneva, @s-kuchi, @summertony15, @troysmullerna, @wagneradrian92, @wally007, @ybizeul, AlessandroN, Didlier, Hubert, Rusty Brow, Tamas Zsolt
:seedling: This release includes 42 features, 38 bug fixes, 10 documentation, 1 performance, 6 styling, 9 refactoring, 16 miscellaneous, and 17 ci pull requests.
:pushpin: Highlights of this major release include:
New Datacenter dashboard which contains node health, capacity, performance, storage efficiency, issues, snapshot, power, and temperature details.
Harvest includes SnapMirror active sync EMS events with alert rules. Thanks @Nikhita-13 for reporting.
Harvest monitors FlexCache performance metrics and includes a new FlexCache dashboard to visualize them. Thanks to @ewilts for raising.
Harvest detects HA pair down and sensor failures. These are shown in the Health dashboard. Thanks to @johnwarlick for raising.
Harvest monitors MetroCluster diagnostics and shows them in the MetroCluster dashboard. Thanks to @wagneradrian92 for reporting.
We improved the performance of all dashboards that include topk queries. Thanks to @mamoep for reporting!
We added filter support for the ZapiPerf collector. See filter for more detail. Thanks to @debbrata-netapp for reporting.
A bin/harvest grafana customize command that writes the dashboards to the filesystem so other programs can manage them. Thanks to @nicolai-hornung-bl for reporting!
We fixed an intermittent latency spike problem that impacted all perf objects. Thanks to @summertony15 and @rodenj1 for reporting this critical issue.
:star: Several of the existing dashboards include new panels in this release:
More Harvest dashboard dropdown variables include the All option. Making it easier to get an overview of your environment.
All EMS alerts include an impact annotation. Thanks to @Divya for raising.
:ear_of_rice: Harvest includes new templates to collect:
:closed_book: Documentation additions
harvest.yml file and share a redacted copy with the Harvest team. Details
All constituents are disabled by default for workload detail performance templates.
The bin/harvest zapi CLI now supports a timeout argument.
Harvest performance collectors (ZapiPerf and RestPerf) ask ONTAP for performance counter metadata every 24 hours instead of every 20 minutes. Thanks to BrianMa for raising.
The Harvest REST collector's api_time metric now includes the API time for all template endpoints. Thanks to ChristopherWilcox for raising.
:bangbang: IMPORTANT Release 24.02 disables four templates that collected metrics not used in dashboards.
These four templates are disabled by default: ObjectStoreClient, TokenManager, OntapS3SVM, and Vscan.
This change was made to reduce the number of collected metrics.
If you require these templates, you can enable them by uncommenting them in their corresponding default.yaml or by extending the existing object template.
:small_red_triangle: IMPORTANT The minimum version of Prometheus required to run Harvest is now 2.33.
Version 2.33 is required to take advantage of Prometheus's @ modifier.
Please upgrade your Prometheus server to at least 2.33 before upgrading Harvest.
:bulb: IMPORTANT After upgrade, don't forget to re-import your dashboards, so you get all the new enhancements and fixes. You can import them via the 'bin/harvest grafana import' CLI, from the Grafana UI, or from the 'Maintenance > Reset Harvest Dashboards' button in NAbox.
Harvest does not calculate power metrics for AFF A250 systems. This data is not available from ONTAP via ZAPI or REST. See ONTAP bug 1511476 for more details.
ONTAP does not include REST metrics for offbox_vscan_server and offbox_vscan until ONTAP 9.13.1. See ONTAP bug
1473892 for more details.
IMPORTANT 7-mode filers that are not on the latest release of ONTAP may experience TLS connection issues with errors
like tls: server selected unsupported protocol version 301 This is caused by a change in Go 1.18.
The default for TLS client connections was changed to TLS 1.2 in Go 1.18.
Please upgrade your 7-mode filers (recommended) or set tls_min_version: tls10 in
your harvest.yml poller section.
See #1007 for more details.
:metal: Thanks to all the people who've opened issues, asked questions on Discord, and contributed code or dashboards this release:
@shawnahall71, @pilot7777, @ben, @Madaan, @johnwarlick, @jfong5040, @santosh725, @summertony15, @jmg011, @cheese1, @mamoep, @Falcon667, @Dess, @debbrata-netapp, @ewilts, @Nikhita-13, @norespers, @nicolai-hornung-bl, @BrendonA667, @schumijo, @Divya, @joshuacook-tamu, @wagneradrian92, @george-strother
:seedling: This release includes 26 features, 24 bug fixes, 20 documentation, 3 styling, 5 refactoring, 11 miscellaneous, and 12 ci pull requests.
Begin Log For Pollcounter And Pollinstance Should Be In Ms (#2509)Storage Efficiency Ratios Panels Should Show Cluster Capacity (#2529)api_times (#2679)Begin Logging To The End Of The Line (#2513):pushpin: Highlights of this major release include:
New FlexGroup dashboard that includes FlexGroup constituents. Thanks to @sandromuc and @ewilts for raising.
Harvest ChangeLog plugin to detect and monitor changes related to object creation, modification, and deletion.
We improved how Harvest calculates power. As a result, you may notice a decrease in the reported power metrics compared to previous versions. Details here. Thanks to Evan Lee for reporting!
Added conf_path variable for specifying the search path of Harvest templates.
:package: Streamlined the Harvest container installation process by eliminating the need to download a tar file. Running Harvest in a container is now simpler and more convenient.
:star: Several of the existing dashboards include new panels in this release:
:ear_of_rice: Harvest includes new templates to collect:
:closed_book: Documentation additions
ca_cert property. Thanks to Marvin Montanus for reporting!Enhanced the Volume dashboard to include clone information.
:zap: Optimized the Harvest binaries, significantly reducing their size.
The Metadata dashboard works inside container deployments.
The FabricPool panels in the Volume dashboard now support FlexGroup volumes. Thanks to @sriniji for reporting.
Large harvest.yml files can be refactoring into smaller ones. Thanks to @llelik and @Pengng88 for raising.
:bulb: Added help text about metrics to more Harvest dashboard panels.
:bangbang: IMPORTANT Due to ONTAP bug 1585893 the Harvest team recommends using ZapiPerf instead of RestPerf when collecting performance metrics. The RestPerf collector can be used once you upgrade your cluster to a version of ONTAP with the fix. Details in 1585893.
:bangbang: IMPORTANT Release 23.11 disables the CIFSSession templates by default. This change was made to prevent the generation of a large number of metrics. If you require these templates, you can enable them. Please be aware that enabling them may result in a significant increase in metric collection time, Harvest memory footprint, and Prometheus used disk space. These metrics are utilized in the SMB2 dashboard.
:bangbang: IMPORTANT Release 23.11 has updated its power metric calculation algorithm. As a result, you may notice a decrease in the reported power metrics compared to previous versions. To collect these metrics, Rest API permissions are required. For detailed information on the power algorithm, please refer to the power algorithm documentation.
:bangbang: IMPORTANT NetApp moved their communities from Slack to Discord, please join us there!
:bulb: IMPORTANT After upgrade, don't forget to re-import your dashboards, so you get all the new enhancements and fixes. You can import them via the bin/harvest grafana import CLI, from the Grafana UI, or from the Maintenance > Reset Harvest Dashboards button in NAbox.
Some AFF A250 systems do not report power metrics. See ONTAP bug 1511476 for more details.
ONTAP does not include REST metrics for offbox_vscan_server and offbox_vscan until ONTAP 9.13.1. See ONTAP bug
1473892 for more details.
IMPORTANT 7-mode filers that are not on the latest release of ONTAP may experience TLS connection issues with errors like tls: server selected unsupported protocol version 301 This is caused by a change in Go 1.18. The default for TLS client connections was changed to TLS 1.2 in Go 1.18. Please upgrade your 7-mode filers (recommended) or set tls_min_version: tls10 in your harvest.yml poller section. See #1007 for more details.
:metal: Thanks to all the people who've opened issues, asked questions on Discord, and contributed code or dashboards this release:
@Garydep, @MrObvious, @Pengng88, @Sandromuc, @ewilts, @faguayot, @jmg011, @kinderr95, @llelik, @mamoep, @rodenj1, @s-kuchi, @shawnahall71, @slater0013, @sriniji, @statdigger, @wyahn1219, AlessandroN, Dave, Diane, Evan Lee, Francesco, Heaven7, Johnathan Warlick, Madaan, Martijn Moret, Marvin Montanus, NicoSeiberth, RBrown, TonyHsieh, Watson9121, dbakerletn, imthenightbird, roller, twodot0h, tymercer
:seedling: This release includes 38 features, 26 bug fixes, 24 documentation, 5 performance, 2 refactoring, 12 miscellaneous, and 7 ci pull requests.
Harvest_conf When Set (#2366)Harvest_conf When Set (#2367)Volume Iops Per Type Legend Should Use Mean, Last, Max Instead… (#2411)Ca_cert Property (#2374)Generate Metrics Should Include Metrics Created By Builtin Plu… (#2413)Conf_path (#2480):pushpin: Highlights of this major release include:
Harvest Security dashboard highlights compliance using NetApp's Security hardening guide for ONTAP
Harvest's credential script supports ONTAP daily credential rotation. Thanks to @mamoep for raising.
:tophat: Harvest makes it easy to run with both the ZAPI and REST collectors at the same time. Overlapping resources are deduplicated and only exported once. Harvest will automatically upgrade ZAPI conversations to REST when ZAPIs are suspended or disabled.
:gem: Updated workload dashboard now includes Service Center, Latency Breakdown, and 50 panels
:gem: Cluster dashboard updated to work with FSx. Some panels are blank because FSx does not have that data.
:mega: The Harvest team published a couple of screencasts about:
:star: Several of the existing dashboards include new panels in this release:
All Harvest dashboard tables include column filters
Harvest dashboards use color to highlight latency and busy threshold breaches
Harvest's Prometheus exporter supports TLS
:ear_of_rice: Harvest includes new templates to collect:
:closed_book: Documentation additions
--labels option of bin/harvest grafana. Thanks to @slater0013 for raisingThe Harvest doctor command validates collector names listed in your harvest.yml file
An earlier version of Harvest collected cloud store information via REST. This release adds the same for ZAPI
When ONTAP resources are missing, Harvest tries to collect them every hour. Earlier versions of Harvest waited 24 hours before retrying, which often caused metrics to be missing after a cluster upgrade. Thanks to @Falcon667 for raising
Earlier versions of Harvest created world writable auto-support files. These files are now only read/writeable by the current user. Thanks to Bunnygirl for raising
bin/harvest import should work with Grafana 10. Thanks to @wooyoungAhn for raising
:bangbang: IMPORTANT 23.08 fixes a REST collector bug that caused partial data collection when ONTAP paginated results. See #2109 for details.
:bangbang: IMPORTANT Release 23.08 disables the NetConnections and NFSClients templates by default. You can enable them if needed. These templates were disabled because several customers reported that these templates created millions of metrics. None of these metrics are used in Harvest dashboards.
:bangbang: IMPORTANT Release 23.08 changes how Harvest monitors workloads. For detailed information, please refer to the discussion #2265.
:bulb: The Compliance dashboard was removed after its panels were moved to the Security dashboard.
:eyes: Ambient temperature metric may experience an increase due to issue #2259
:bangbang: IMPORTANT NetApp moved their communities from Slack to Discord, please join us there!
:bangbang: IMPORTANT If using Docker Compose and you want to keep your historical Prometheus data, please read how to migrate your Prometheus volume
:bulb: IMPORTANT After upgrade, don't forget to re-import your dashboards, so you get all the new enhancements and fixes. You can import them via the bin/harvest grafana import CLI, from the Grafana UI, or from the Maintenance > Reset Harvest Dashboards button in NAbox.
Some AFF A250 systems do not report power metrics. See ONTAP bug 1511476 for more details.
ONTAP does not include REST metrics for offbox_vscan_server and offbox_vscan until ONTAP 9.13.1. See ONTAP bug
1473892 for more details.
IMPORTANT 7-mode filers that are not on the latest release of ONTAP may experience TLS connection issues with errors like tls: server selected unsupported protocol version 301 This is caused by a change in Go 1.18. The default for TLS client connections was changed to TLS 1.2 in Go 1.18. Please upgrade your 7-mode filers (recommended) or set tls_min_version: tls10 in your harvest.yml poller section. See #1007 for more details.
:metal: Thanks to all the people who've opened issues, asked questions on Discord, and contributed code or dashboards this release:
@7840vz, @DAx-cGn, @Falcon667, @Hedius, @LukaszWasko, @MrObvious, @ReneMeier, @Sawall10, @T1r0l, @XDavidT, @amd-eulee, @aticatac, @chadpruden, @cwaltham, @cygio, @ddhti, @debert-ntap, @demalik, @electrocreative, @elsgaard, @ev1963, @faguayot, @iStep2Step, @jgasher, @jmg011, @lobster1860, @mamoep, @matejzero, @matthieu-sudo, @merdos, @pilot7777, @rbrownATnetapp, @rodenj1, @slater0013, @swordfish291, @tsohst, @wooyoungAhn, Alessandro.Nuzzo, Ed Wilts, GregS, Imthenightbird, KlausHub, MeghanaD, MikeK, Paul P2, Rusty Brown, Shubham Mer, Tudor Pascu, Watson9121, jf38800, jfong, lorenzoc, rcl23, roller, scrhobbs, troysmuller, twodot0h
:seedling: This release includes 42 features, 40 bug fixes, 20 documentation, 2 performance, 4 testing, 1 styling, 9 refactoring, 20 miscellaneous, and 12 ci pull requests.
Go Build Command To Build Harvest And Poller Binaries (#2221)Seconds Unit (#2039)-Logtofile Option (#2119)Addr Argument To Credentials_script (#2128):pushpin: Highlights of this major release include:
:gem: Seven new dashboards:
:star: Several of the existing dashboards include new panels in this release:
:closed_lock_with_key: Harvest can fetch cluster credentials via a credential script. Thanks to Ed Wilts for raising.
:ear_of_rice: Harvest includes new templates to collect:
Harvest containers are published to GitHub's container registry in addition to DockerHub and cr.netapp.io.
If you're using cr.netapp.io, we encourage you to switch to ghcr.io or DockerHub. In 2024, we will stop publishing to cr.netapp.io
Harvest uses a distroless image as its base now - reducing the size of the container and reducing the attack surface
Harvest collects 38 additional EMS events and alert rules in this release
Harvest EMS alert rules were updated to include better label names and align their severity with Prometheus best practices. Thanks to @7840vz for contributing this feature!
The bin/harvest doctor tool validates your custom.yaml template files, checking them for errors.
:closed_book: Documentation additions
prefer_zapi flag that tells Harvest to use the ZAPI API if the cluster supports it, otherwise allow Harvest to choose REST or ZAPI, whichever is appropriate to the ONTAP version. See rest-strategy for details.:tophat: Harvest makes it easy to run with both the ZAPI and REST collectors at the same time. Overlapping resources are deduplicated and only published to Prometheus once. This was the final piece in our journey to REST. See rest-strategy.md if you are interested in the details.
IMPORTANT The volume_aggr_labels metric is being deprecated in the 23.05 release and will be removed in the 23.08 release of Harvest (#1966) volume_aggr_labels is redundant and the same labels are already available via volume_labels.
IMPORTANT To reduce image and download size, several tools were combined in 23.05. The following binaries are no longer included: bin/grafana, bin/rest, bin/zapi. Use bin/harvest grafana, bin/harvest rest, and bin/harvest zapi instead.
:bangbang: IMPORTANT NetApp moved their communities from Slack to Discord, please join us there!
:bangbang: IMPORTANT If using Docker Compose and you want to keep your historical Prometheus data, please read how to migrate your Prometheus volume
:bulb: IMPORTANT After upgrade, don't forget to re-import your dashboards, so you get all the new enhancements and fixes. You can import them via the bin/harvest grafana import CLI, from the Grafana UI, or from the Maintenance > Reset Harvest Dashboards button in NAbox.
Harvest does not calculate power metrics for AFF A250 systems. This data is not available from ONTAP via ZAPI or REST. See ONTAP bug 1511476 for more details.
ONTAP does not include REST metrics for offbox_vscan_server and offbox_vscan until ONTAP 9.13.1. See ONTAP bug
1473892 for more details.
IMPORTANT 7-mode filers that are not on the latest release of ONTAP may experience TLS connection issues with errors like tls: server selected unsupported protocol version 301 This is caused by a change in Go 1.18. The default for TLS client connections was changed to TLS 1.2 in Go 1.18. Please upgrade your 7-mode filers (recommended) or set tls_min_version: tls10 in your harvest.yml poller section. See #1007 for more details.
:metal: Thanks to all the people who've opened issues, asked questions on Discord, and contributed code or dashboards this release:
@7840vz, @DAx-cGn, @Falcon667, @Hedius, @LukaszWasko, @MrObvious, @ReneMeier, @Sawall10, @T1r0l, @XDavidT, @aticatac, @chadpruden, @cygio, @ddhti, @debert-ntap, @demalik, @electrocreative, @elsgaard, @ev1963, @faguayot, @iStep2Step, @jgasher, @jmg011, @mamoep, @matejzero, @matthieu-sudo, @merdos, @pilot7777, @rodenj1, Alessandro.Nuzzo, Ed Wilts, Imthenightbird, KlausHub, MeghanaD, Paul P2, Rusty Brown, Shubham Mer, Tudor Pascu, Watson9121, jf38800, jfong, rcl23, troysmuller, twodot0h
:seedling: This release includes 61 features, 49 bug fixes, 22 documentation, 2 testing, 8 refactoring, 25 miscellaneous, and 32 ci pull requests.
--Volume Option For Additio… (#1924)Node_nfs_latency Is Microsecs (#1853)--Port By Default (#1917)Validateportinuse That Caused Test To Fail (#1889):pushpin: Highlights of this major release include:
:sparkles: Harvest includes a new file system analytics (FSA) dashboard with directory growth, top directories per volume, and volume usage statistics.
Harvest includes a new StorageGRID overview dashboard with performance, storage, information lifecycle management, and node panels. We're collecting suggestions on which StorageGRID dashboards you'd like to see next in issue #1420.
:bulb: Power dashboard includes new panels for total power by aggregate disk type, average power per used TB, average IOPs/Watt, total power by aggregate, and information on sensor problems.
:tophat: Harvest makes it easy to run with both the ZAPI and REST collectors at the same time. Overlapping resources are deduplicated and only published to Prometheus once. This was the final piece in our journey to REST. See rest-strategy.md if you are interested in the details.
:closed_book: We made lots of improvements to Harvest's new documentation site this release including one of the most requested features - a list of Harvest metrics and their corresponding ONTAP ZAPI/REST API mappings. :triangular_ruler: Check it out
:gem: New dashboards and improvements
Harvest includes a new command bin/harvest grafana metrics which shows which metrics each dashboard uses
:bangbang: IMPORTANT NetApp moved their communities from Slack to Discord, please join us there!
:bangbang: IMPORTANT If using Docker Compose and you want to keep your historical Prometheus data, please read how to migrate your Prometheus volume
:bulb: IMPORTANT After upgrade, don't forget to re-import your dashboards, so you get all the new enhancements and
fixes. You can import them via the bin/harvest/grafana import CLI, from the Grafana UI, or from
the Maintenance > Reset Harvest Dashboards button in NAbox.
:sunflower: In the 22.11.0 release notes, we announced that we would be removing quota metrics prefixed with qtree.
Several of you asked us to leave them. :+1: We will continue publishing them as-is.
Harvest does not calculate power metrics for AFF A250 systems. This data is not available from ONTAP via ZAPI or REST. See ONTAP bug 1511476 for more details.
ONTAP does not include REST metrics for offbox_vscan_server and offbox_vscan until ONTAP 9.13.1. See ONTAP bug
1473892 for more details.
IMPORTANT 7-mode filers that are not on the latest release of ONTAP may experience TLS connection issues with errors
like tls: server selected unsupported protocol version 301 This is caused by a change in Go 1.18. The default for TLS client connections was changed to TLS 1.2 in Go 1.18. Please upgrade your 7-mode filers (recommended) or set tls_min_version: tls10 in your harvest.yml poller section. See #1007 for more details.
:metal: Thanks to all the people who've opened issues, asked questions on Discord, and contributed code or dashboards this release:
@Falcon667, @MrObvious, @ReneMeier, @Sawall10, @T1r0l, @aticatac, @chadpruden, @demalik, @electrocreative, @ev1963, @faguayot, @iStep2Step, @jgasher, @jmg011, @mamoep, @matejzero, @matthieu-sudo, @merdos, @rodenj1, Ed Wilts, KlausHub, MeghanaD, Paul P2, Rusty Brown, Shubham Mer, Tudor Pascu, jf38800, jfong, rcl23, troysmuller, twodot0h
:seedling: This release includes 43 features, 43 bug fixes, 19 documentation, 2 testing, 1 styling, 5 miscellaneous, and 7 ci pull requests.
Bin/Grafana Metrics To Print Which Metrics Each Dashboard… (#1578)--Overwrite Does When Importing Dashboards That … (#1574)Metrics Query Permissions For Storagegrid (#1687)Status/Testme (#1541)Status/Testme (#1552):pushpin: This release is the same as 22.11.0 with a fix that makes the REST Perf collector work with the just released ONTAP 9.12.1+ and a few other minor fixes. If you are not monitoring ONTAP 9.12.1, you can ignore this release.
Full Changelog: https://github.com/NetApp/harvest/compare/v22.11.0...v22.11.1
:pushpin: Highlights of this major release include:
:sparkles: Harvest now includes a StorageGRID collector and a Tenant/Buckets dashboard. We're just getting started with StorageGRID dashboards. Please give the collector a try, and let us know which StorageGRID dashboards you'd like to see next.
:tophat: The REST collectors are ready! We recommend using them for ONTAP versions 9.12.1 and higher. Today, Harvest collects 1,546 metrics via ZAPI. Harvest includes a full set of REST templates that export identical metrics. All 1,546 metrics are available via Harvest's REST templates and no changes to dashboards or downstream metric-consumers is required. :tada: More details on Harvest's REST strategy.
:closed_book: Harvest has a new documentation site! This consolidates Harvest documentation into one place and will make it easier to find what you need. Stay tuned for more updates here.
:gem: New and improved dashboards
:lock: Harvest code is checked for vulnerabilities on every commit using Go's vulnerability management scanner.
Harvest collects additional metrics in this release
Harvest now converts ONTAP histograms to Prometheus histograms, making it possible to visualize metrics as heatmaps in Grafana
:bangbang: IMPORTANT NetApp moved their communities from Slack to Discord, please join us there!
:bomb: Deprecation: Earlier versions of Harvest published quota metrics prefixed with qtree. Harvest release 22.11 deprecates the quota metrics prefixed with qtree and instead publishes quota metrics prefixed with quota. All dashboards have been updated. If you are consuming these metrics outside the default dashboards, please change to quota prefixed metrics. Harvest release 23.02 will remove the deprecated quota metrics prefixed with qtree.
:bangbang: IMPORTANT If using Docker Compose and you want to keep your historical Prometheus data, please read how to migrate your Prometheus volume
:bulb: IMPORTANT After upgrade, don't forget to re-import your dashboards, so you get all the new enhancements and fixes. You can import them via the bin/harvest/grafana import CLI, from the Grafana UI, or from the Maintenance > Reset Harvest Dashboards button in NAbox.
Harvest does not calculate power metrics for AFF A250 systems. This data is not available from ONTAP via ZAPI or REST. See ONTAP bug 1511476 for more details.
ONTAP does not include REST metrics for offbox_vscan_server and offbox_vscan until ONTAP 9.13.1. See ONTAP bug 1473892 for more details.
Podman is unable to pull from NetApp's container registry cr.netapp.io. Until issue is resolved, Podman users can pull from a separate proxy like this podman pull netappdownloads.jfrog.io/oss-docker-harvest-production/harvest:latest.
7-mode filers that are not on the latest release of ONTAP may experience TLS connection issues with errors like tls: server selected unsupported protocol version 301 This is caused by a change in Go 1.18. The default for TLS client connections was changed to TLS 1.2 in Go 1.18. Please upgrade your 7-mode filers (recommended) or set tls_min_version: tls10 in your harvest.yml poller section. See #1007 for more details.
The Unix collector is unable to monitor pollers running in containers. See #249 for details.
:metal: Thanks to all the people who've opened issues, asked questions on Discord, and contributed code or dashboards this release:
@Falcon667, @MrObvious, @ReneMeier, @Sawall10, @T1r0l, @chadpruden, @demalik, @electrocreative, @ev1963, @faguayot, @iStep2Step, @jgasher, @jmg011, @mamoep, @matthieu-sudo, @merdos, @rodenj1, Ed Wilts, KlausHub, MeghanaD, Paul P2, Rusty Brown, Shubham Mer, jf38800, rcl23, troysmuller
:seedling: This release includes 59 features, 90 bug fixes, 21 documentation, 4 testing, 2 styling, 6 refactoring, 2 miscellaneous, and 6 ci commits.
Make Dev (#1273)Bin/Zapi Show Counters Should Print Xml Results To Make Parsi… (#1286)Objects And Collectors Override Defaults (#1300)Shared Crosshair (#1359)Harvest_no_upgrade To Skip Collector Upgrade (#1385)Defer Resp.body.close() Will Not Run (#1211)Targetisontap (#1235)