From bbe6bd6a7a1b18aca3ea4ad91beb8440e99854ec Mon Sep 17 00:00:00 2001 From: Brandon Morelli Date: Wed, 29 May 2019 11:46:56 -0700 Subject: [PATCH 1/5] docs: ILM first draft --- .../command-reference.asciidoc | 7 + docs/ilm.asciidoc | 187 ++++++++++++++++++ 2 files changed, 194 insertions(+) create mode 100644 docs/ilm.asciidoc diff --git a/docs/copied-from-beats/command-reference.asciidoc b/docs/copied-from-beats/command-reference.asciidoc index e1ed7c3146f..d456c595ce8 100644 --- a/docs/copied-from-beats/command-reference.asciidoc +++ b/docs/copied-from-beats/command-reference.asciidoc @@ -636,8 +636,15 @@ enabled modules in the +{beatname_lc}.yml+ file. If you used the directory, also specify the `--modules` flag. endif::[] +ifndef::apm-server[] *`--template`*:: Sets up the index template only. +endif::apm-server[] + +ifdef::apm-server[] +*` --index-management`*:: +Sets up <>. +endif::[] {global-flags} diff --git a/docs/ilm.asciidoc b/docs/ilm.asciidoc new file mode 100644 index 00000000000..056f246222e --- /dev/null +++ b/docs/ilm.asciidoc @@ -0,0 +1,187 @@ +[[ilm]] +[role="xpack"] +== Index lifecycle management (ILM) + +Use the {ref}/getting-started-index-lifecycle-management.html[index lifecycle management] +feature in {es} to manage your APM Server indices as they age. +For example, instead of creating daily indices where index size can vary based on the number of agents and number of events sent, +use an index lifecycle policy to automate a rollover to a new index when the existing index reaches a specified size or age. + +You can view and edit the index lifecycle policies in the *Index lifecycle policies* UI in {kib}. +For more information about working with the UI, +see {kibana-ref}/index-lifecycle-policies.html[Index lifecyle policies]. + +NOTE: If you're migrating from an existing setup, +any indices present before ILM was configured will need to be managed manually. + +[float] +[[ilm-setup]] +=== ILM default configuration + +IMPORTANT: The default index lifecycle configuration cannot be configured until after it has been applied. + +The default configuration only applies hot and warm phase definitions. +Cold and delete phases are not defined. +Because errors and spans lose information value faster than metrics and transactions do, +there are two different policies defined: +one for `errors` and `spans`, and one for `metrics` and `transactions`. + +[options="header"] +|======================================================================= +|Event type |Hot |Warm +|Errors & Spans +|Rollover: `max_size: 50gb`, `max_age: 1 day` +`priority: 100` +|`min_age: 7 days` +readonly +`priority: 50` +|Transactions & Metrics +|Rollover: `max_size: 50gb`, `max_age: 7 days` +`priority: 100` +|`min_age: 31 days` +readonly +`priority: 50` +|======================================================================= + +After applying ILM, the policies are editable in the Kibana **UI management** section. + +// OLD INFO, MAYBE DELETE +// ******* + +**Error & Span events** +_14 days to look at data in a very fast way_ +HOT +* A new index is created every 50gb or 1 day (whichever comes first). This makes it very easy to delete indices day by day, and ensures no index gets too large. +* After seven days, hot indices are moved to the warm phase. +WARM +* Warm indices are read-only. +* After another seven days, warm indices are moved to the cold phase. +COLD +* Cold indices move data out of memory to clean up memory usage. Data can still be queried, albeit slowly. +DELETE +* By default, no delete phase is defined. Create a custom delete policy using the UI management section. + +**Transactions & Metrics** +_Compare the current month with your last month without any performance reduction_ +HOT +* A new index is created every 50gb or 7 days (whichever comes first). This makes it very easy to delete indices week by week, and ensures no index gets too large. +* After one month (31 days), hot indices are moved to the warm phase. +WARM +* Warm indices are read-only. +* After two more months (62 days), warm indices are moved to the cold phase. +COLD +* Cold indices move data out of memory to clean up memory usage. Data can still be queried, albeit slowly. +DELETE +* By default, no delete phase is defined. Create a custom delete policy using the UI management section. + +// ******* +// END OLD INFO + +[float] +[[ilm-setup]] +=== ILM setup + +To use index lifecylce management, set `ilm.enabled` to `true` in apm-server.yml. + +ILM can only be enabled for `output.elasticsearch`. +When enabled, configurations defined for `output.elasticsearch.index` and `output.elasticsearch.indices` will be ignored. +ILM clears all index settings and sets up a predefined set of indices and policies that the user cannot configure. + +It is recommended to setup index lifecycle management (ILM) before starting APM Server. +This prevents indices from being created before policies are changed, and allows policies to be applied to indices at creation time. +It also excludes setup from the ingestion process, which allows you to ensure ILM is setup correctly before using APM. + +Index lifecycle management will manage an index based on its defined policy. +Run the <> with the ` --index-management` flag to set up the default ILM policy: + +[source,js] +----------------------- +./apm-server setup --index-management +----------------------- +// CONSOLE + +Policies only need to be created once, and will persist through version upgrades. +You can confirm the policy was created with the GET lifecycle policy API. +Here's what the transaction response looks like: + +[source,js] +----------------------- +GET _ilm/policy +{ + "apm-7.2.0-transaction": { + "version": 1, + "modified_date": "2019-05-28T15:55:26.791Z", + "policy": { + "phases": { + "warm": { + "min_age": "31d",<3> + "actions": { + "readonly": {}, + "set_priority": { + "priority": 50<4> + } + } + }, + "hot": { + "min_age": "0ms", + "actions": { + "rollover": { + "max_size": "50gb",<1> + "max_age": "7d"<2> + }, + "set_priority": { + "priority": 100<4> + } + } + } + } + } + } +} +----------------------- +// CONSOLE +<1> Rollover after _50gb_ +<2> Rollover after _7 days_ +<3> Move to warm phase after _31 days_ +<4> Priority for recovering your indices after a node restart. Higher priorities are recovered first. + +[float] +[[ilm-edit]] +=== Editing ILM defaults + +Once you've enabled index lifecycle management and setup the ILM policies, you can view and edit the policies in the Index lifecycle policies UI in Kibana. For more information about working with the UI, +see {kibana-ref}/index-lifecycle-policies.html[Index lifecyle policies]. + + +// NOTES BELOW... +// ************** + +// True?? If you do not run the setup command, policies will will not be applied immediately?? + +// We want users to have the option to define policies and prefixes for the indices, not the whole config. + +When you start the APM Server, you'll see in the logs index-management. +The template gets set up the same as before. +The ILM policy gets setup. +Load template error-ilm template (for every type of event) template, policy, alias + +--- +GET _template/apm* +--- + +Used to just be the `apm-{observer.version}` template. +Now, there are four additional templates, one for each event. Each has an order (`"order": 2`). If multiple templates contain the same key, ES prefers the template with the lower number. Each event template sets the index lifecycle. + +We prefix policies and templates with version numbers. This means if you use apm server, then modify your policy and do an upgrade, you'll need to do two things: +1. They set ILM enabled to false and configure the indices to the old ILM indices in yml. Then, no new policy +2. They need to update the policies manually again when they upgrade. When APM Server is started for the first time, APM Server checks if the policies are there for this version. Update custom rollover time/size. + +Updating a policy wont take effect until a new index is created. Make sure you update all ILM settings prior to starting APM Server if you are not OK with the defaults. + +// Questions: +Global Kibana setting for in/excluding frozen indices in searches. If set to false, frozen data will not be queried. +Changing this global setting could potentially slow down other queries. +https://www.elastic.co/guide/en/elasticsearch/reference/7.x/multi-index.html +https://www.elastic.co/guide/en/elasticsearch/reference/7.x/searching_a_frozen_index.html + +Do we remove manual ILM setup? \ No newline at end of file From 7a71e4873078965b80f18dc23731cbd11e3feb40 Mon Sep 17 00:00:00 2001 From: Brandon Morelli Date: Wed, 29 May 2019 12:16:54 -0700 Subject: [PATCH 2/5] docs: ILM updates --- docs/configuring.asciidoc | 2 +- docs/ilm.asciidoc | 29 ++++++++++++++++------------- 2 files changed, 17 insertions(+), 14 deletions(-) diff --git a/docs/configuring.asciidoc b/docs/configuring.asciidoc index 0def3f96690..390de5b4c97 100644 --- a/docs/configuring.asciidoc +++ b/docs/configuring.asciidoc @@ -27,7 +27,7 @@ include::./copied-from-beats/shared-ssl-config.asciidoc[] include::./template-config.asciidoc[] -include::./ilm-setup.asciidoc[] +include::./ilm.asciidoc[] include::./copied-from-beats/loggingconfig.asciidoc[] diff --git a/docs/ilm.asciidoc b/docs/ilm.asciidoc index 056f246222e..cc15f8269a2 100644 --- a/docs/ilm.asciidoc +++ b/docs/ilm.asciidoc @@ -18,7 +18,7 @@ any indices present before ILM was configured will need to be managed manually. [[ilm-setup]] === ILM default configuration -IMPORTANT: The default index lifecycle configuration cannot be configured until after it has been applied. +IMPORTANT: The default index lifecycle configuration cannot be changed until after it has been applied. The default configuration only applies hot and warm phase definitions. Cold and delete phases are not defined. @@ -43,8 +43,6 @@ readonly `priority: 50` |======================================================================= -After applying ILM, the policies are editable in the Kibana **UI management** section. - // OLD INFO, MAYBE DELETE // ******* @@ -81,11 +79,10 @@ DELETE [[ilm-setup]] === ILM setup -To use index lifecylce management, set `ilm.enabled` to `true` in apm-server.yml. - +To set up index lifecycle management, set `ilm.enabled` to `true` in apm-server.yml. ILM can only be enabled for `output.elasticsearch`. When enabled, configurations defined for `output.elasticsearch.index` and `output.elasticsearch.indices` will be ignored. -ILM clears all index settings and sets up a predefined set of indices and policies that the user cannot configure. +ILM clears all index settings and sets up a predefined set of indices and policies that the user cannot initially configure. It is recommended to setup index lifecycle management (ILM) before starting APM Server. This prevents indices from being created before policies are changed, and allows policies to be applied to indices at creation time. @@ -100,7 +97,7 @@ Run the <> with the ` --index-management` flag to set up the defa ----------------------- // CONSOLE -Policies only need to be created once, and will persist through version upgrades. +//Policies only need to be created once, and will persist through version upgrades. You can confirm the policy was created with the GET lifecycle policy API. Here's what the transaction response looks like: @@ -150,7 +147,13 @@ GET _ilm/policy === Editing ILM defaults Once you've enabled index lifecycle management and setup the ILM policies, you can view and edit the policies in the Index lifecycle policies UI in Kibana. For more information about working with the UI, -see {kibana-ref}/index-lifecycle-policies.html[Index lifecyle policies]. +see {kibana-ref}/index-lifecycle-policies.html[*Index lifecycle policies*]. + +Index policies and templates are prefixed with version numbers, for example: `apm-7.2.0-metric-000001`. +This means if you use APM Server, modify the policy, and upgrade, you'll need to do two things: +1. Set `ilm.enabled` to false and configure your indices to the old ILM indices in the yml file. +2. Update the policies manually when you upgrade. When APM Server is started for the first time, +APM Server checks to see if the policies are there for this version, if they're not, it creates them. // NOTES BELOW... @@ -172,10 +175,6 @@ GET _template/apm* Used to just be the `apm-{observer.version}` template. Now, there are four additional templates, one for each event. Each has an order (`"order": 2`). If multiple templates contain the same key, ES prefers the template with the lower number. Each event template sets the index lifecycle. -We prefix policies and templates with version numbers. This means if you use apm server, then modify your policy and do an upgrade, you'll need to do two things: -1. They set ILM enabled to false and configure the indices to the old ILM indices in yml. Then, no new policy -2. They need to update the policies manually again when they upgrade. When APM Server is started for the first time, APM Server checks if the policies are there for this version. Update custom rollover time/size. - Updating a policy wont take effect until a new index is created. Make sure you update all ILM settings prior to starting APM Server if you are not OK with the defaults. // Questions: @@ -184,4 +183,8 @@ Changing this global setting could potentially slow down other queries. https://www.elastic.co/guide/en/elasticsearch/reference/7.x/multi-index.html https://www.elastic.co/guide/en/elasticsearch/reference/7.x/searching_a_frozen_index.html -Do we remove manual ILM setup? \ No newline at end of file +Do we remove manual ILM setup? + +Privileges required for APM Server: +cluster: manage_ilm +indices: manage, manage_ilm \ No newline at end of file From 9ec950243a78e5e097aed595cf0a1d1f05f277df Mon Sep 17 00:00:00 2001 From: Brandon Morelli Date: Tue, 4 Jun 2019 10:36:31 -0700 Subject: [PATCH 3/5] docs: ILM feedback and improvements --- docs/guide/apm-release-notes.asciidoc | 11 ++ docs/ilm-setup.asciidoc | 10 +- docs/ilm.asciidoc | 142 +++++++------------------- 3 files changed, 53 insertions(+), 110 deletions(-) diff --git a/docs/guide/apm-release-notes.asciidoc b/docs/guide/apm-release-notes.asciidoc index 3c9bb13eb7a..b31be3baafc 100644 --- a/docs/guide/apm-release-notes.asciidoc +++ b/docs/guide/apm-release-notes.asciidoc @@ -7,6 +7,7 @@ For a full list of changes, see the {apm-server-ref-v}/release-notes.html[APM Server Release Notes] or the {kibana-ref}/release-notes.html[Kibana Release Notes]. +* <> * <> * <> * <> @@ -16,6 +17,16 @@ For a full list of changes, see the * <> * <> +[[release-highlights-7.2.0]] +=== APM version 7.2.0 + +[float] +==== New features + +*APM Server* + +* Support for {apm-overview-ref-v}/ilm.html[index lifecycle management (ILM)] + [[release-highlights-7.1.0]] === APM version 7.1.0 diff --git a/docs/ilm-setup.asciidoc b/docs/ilm-setup.asciidoc index 231dc3ab7ee..6c80e267f4b 100644 --- a/docs/ilm-setup.asciidoc +++ b/docs/ilm-setup.asciidoc @@ -1,13 +1,9 @@ [role="xpack"] [[manual-ilm-setup]] -== Manual index lifecycle management +=== Manual index lifecycle management -APM Server can take advantage of the index lifecycle management (ILM) capabilities of Elasticsearch. -ILM enables you to automate how you want to manage your indices over time. -You can base actions on factors such as shard size and performance requirements. - -The guide below will help you set up a custom ILM policy for span indices. -You can repeat the actions for any other indices. +The guide below will help you set up a custom index lifecycle management policy for span indices. +You can repeat these actions for any other indices (transactions, errors, etc.). NOTE: If you're migrating from an existing setup, any indices present before ILM was configured will need to be managed manually. diff --git a/docs/ilm.asciidoc b/docs/ilm.asciidoc index cc15f8269a2..0798c0866b5 100644 --- a/docs/ilm.asciidoc +++ b/docs/ilm.asciidoc @@ -2,25 +2,28 @@ [role="xpack"] == Index lifecycle management (ILM) -Use the {ref}/getting-started-index-lifecycle-management.html[index lifecycle management] +Use the {ref}/getting-started-index-lifecycle-management.html[index lifecycle management (ILM)] feature in {es} to manage your APM Server indices as they age. -For example, instead of creating daily indices where index size can vary based on the number of agents and number of events sent, -use an index lifecycle policy to automate a rollover to a new index when the existing index reaches a specified size or age. +ILM enables you to automate how you want to manage your indices over time, +by automating rollovers to a new index when the existing index reaches a specified size or age. You can view and edit the index lifecycle policies in the *Index lifecycle policies* UI in {kib}. For more information about working with the UI, see {kibana-ref}/index-lifecycle-policies.html[Index lifecyle policies]. +APM Server currently offers two ways to get started with ILM: + +1. <> - Get up and running with a default index lifecycle management policy as quickly as possible. +2. <> - Customize and manage your own ILM policies. + NOTE: If you're migrating from an existing setup, any indices present before ILM was configured will need to be managed manually. -[float] -[[ilm-setup]] -=== ILM default configuration - -IMPORTANT: The default index lifecycle configuration cannot be changed until after it has been applied. +[[ilm-default]] +=== ILM default policy -The default configuration only applies hot and warm phase definitions. +Index lifecycle management will manage an index based on its defined policy. +The default ILM configuration applies hot and warm phase policies. Cold and delete phases are not defined. Because errors and spans lose information value faster than metrics and transactions do, there are two different policies defined: @@ -30,65 +33,33 @@ one for `errors` and `spans`, and one for `metrics` and `transactions`. |======================================================================= |Event type |Hot |Warm |Errors & Spans -|Rollover: `max_size: 50gb`, `max_age: 1 day` +|Rollover: `max_size: 50gb`, `max_age: 1 day`, `priority: 100` -|`min_age: 7 days` -readonly +|`min_age: 7 days`, +readonly, `priority: 50` |Transactions & Metrics -|Rollover: `max_size: 50gb`, `max_age: 7 days` +|Rollover: `max_size: 50gb`, `max_age: 7 days`, `priority: 100` -|`min_age: 31 days` -readonly +|`min_age: 31 days`, +readonly, `priority: 50` |======================================================================= -// OLD INFO, MAYBE DELETE -// ******* - -**Error & Span events** -_14 days to look at data in a very fast way_ -HOT -* A new index is created every 50gb or 1 day (whichever comes first). This makes it very easy to delete indices day by day, and ensures no index gets too large. -* After seven days, hot indices are moved to the warm phase. -WARM -* Warm indices are read-only. -* After another seven days, warm indices are moved to the cold phase. -COLD -* Cold indices move data out of memory to clean up memory usage. Data can still be queried, albeit slowly. -DELETE -* By default, no delete phase is defined. Create a custom delete policy using the UI management section. - -**Transactions & Metrics** -_Compare the current month with your last month without any performance reduction_ -HOT -* A new index is created every 50gb or 7 days (whichever comes first). This makes it very easy to delete indices week by week, and ensures no index gets too large. -* After one month (31 days), hot indices are moved to the warm phase. -WARM -* Warm indices are read-only. -* After two more months (62 days), warm indices are moved to the cold phase. -COLD -* Cold indices move data out of memory to clean up memory usage. Data can still be queried, albeit slowly. -DELETE -* By default, no delete phase is defined. Create a custom delete policy using the UI management section. - -// ******* -// END OLD INFO +IMPORTANT: Changes to the default index lifecycle policy do not take effect until the current index has rolled over. +If you'd like to manage a custom policy, see <>. [float] [[ilm-setup]] -=== ILM setup +==== ILM default policy setup To set up index lifecycle management, set `ilm.enabled` to `true` in apm-server.yml. ILM can only be enabled for `output.elasticsearch`. When enabled, configurations defined for `output.elasticsearch.index` and `output.elasticsearch.indices` will be ignored. -ILM clears all index settings and sets up a predefined set of indices and policies that the user cannot initially configure. -It is recommended to setup index lifecycle management (ILM) before starting APM Server. -This prevents indices from being created before policies are changed, and allows policies to be applied to indices at creation time. -It also excludes setup from the ingestion process, which allows you to ensure ILM is setup correctly before using APM. +It is recommended to set up index lifecycle management (ILM) before starting APM Server. +This excludes setup from the ingestion process, which allows you to ensure ILM is set up correctly before using APM. -Index lifecycle management will manage an index based on its defined policy. Run the <> with the ` --index-management` flag to set up the default ILM policy: [source,js] @@ -97,7 +68,6 @@ Run the <> with the ` --index-management` flag to set up the defa ----------------------- // CONSOLE -//Policies only need to be created once, and will persist through version upgrades. You can confirm the policy was created with the GET lifecycle policy API. Here's what the transaction response looks like: @@ -111,11 +81,11 @@ GET _ilm/policy "policy": { "phases": { "warm": { - "min_age": "31d",<3> + "min_age": "31d",<1> "actions": { "readonly": {}, "set_priority": { - "priority": 50<4> + "priority": 50 } } }, @@ -123,8 +93,8 @@ GET _ilm/policy "min_age": "0ms", "actions": { "rollover": { - "max_size": "50gb",<1> - "max_age": "7d"<2> + "max_size": "50gb",<2> + "max_age": "7d"<3> }, "set_priority": { "priority": 100<4> @@ -137,54 +107,20 @@ GET _ilm/policy } ----------------------- // CONSOLE -<1> Rollover after _50gb_ -<2> Rollover after _7 days_ -<3> Move to warm phase after _31 days_ +<1> Move to warm phase after _31 days_ +<2> Rollover after _50gb_ +<3> Rollover after _7 days_ <4> Priority for recovering your indices after a node restart. Higher priorities are recovered first. -[float] -[[ilm-edit]] -=== Editing ILM defaults - -Once you've enabled index lifecycle management and setup the ILM policies, you can view and edit the policies in the Index lifecycle policies UI in Kibana. For more information about working with the UI, -see {kibana-ref}/index-lifecycle-policies.html[*Index lifecycle policies*]. - -Index policies and templates are prefixed with version numbers, for example: `apm-7.2.0-metric-000001`. -This means if you use APM Server, modify the policy, and upgrade, you'll need to do two things: -1. Set `ilm.enabled` to false and configure your indices to the old ILM indices in the yml file. -2. Update the policies manually when you upgrade. When APM Server is started for the first time, -APM Server checks to see if the policies are there for this version, if they're not, it creates them. - - -// NOTES BELOW... -// ************** - -// True?? If you do not run the setup command, policies will will not be applied immediately?? - -// We want users to have the option to define policies and prefixes for the indices, not the whole config. - -When you start the APM Server, you'll see in the logs index-management. -The template gets set up the same as before. -The ILM policy gets setup. -Load template error-ilm template (for every type of event) template, policy, alias - ---- -GET _template/apm* ---- - -Used to just be the `apm-{observer.version}` template. -Now, there are four additional templates, one for each event. Each has an order (`"order": 2`). If multiple templates contain the same key, ES prefers the template with the lower number. Each event template sets the index lifecycle. - -Updating a policy wont take effect until a new index is created. Make sure you update all ILM settings prior to starting APM Server if you are not OK with the defaults. +Your indices are now configured to use index lifecylce management. Go ahead and <>. -// Questions: -Global Kibana setting for in/excluding frozen indices in searches. If set to false, frozen data will not be queried. -Changing this global setting could potentially slow down other queries. -https://www.elastic.co/guide/en/elasticsearch/reference/7.x/multi-index.html -https://www.elastic.co/guide/en/elasticsearch/reference/7.x/searching_a_frozen_index.html +// [float] +// [[ilm-upgrade]] +// ==== ILM default policy upgrades -Do we remove manual ILM setup? +// Policies are versioned so they can change. +// Indices are versioned so they can change. +// An upgrade creates new templates, policies, and indices. +// If you customize anything, it will revert back to the default during an upgrade -Privileges required for APM Server: -cluster: manage_ilm -indices: manage, manage_ilm \ No newline at end of file +include::./ilm-setup.asciidoc[] \ No newline at end of file From f68dc33b01ce15a60559667d5330398bdc081b2c Mon Sep 17 00:00:00 2001 From: Brandon Morelli Date: Tue, 4 Jun 2019 11:14:43 -0700 Subject: [PATCH 4/5] docs: link fix --- docs/guide/apm-release-notes.asciidoc | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/guide/apm-release-notes.asciidoc b/docs/guide/apm-release-notes.asciidoc index b31be3baafc..73be25c1302 100644 --- a/docs/guide/apm-release-notes.asciidoc +++ b/docs/guide/apm-release-notes.asciidoc @@ -25,7 +25,7 @@ For a full list of changes, see the *APM Server* -* Support for {apm-overview-ref-v}/ilm.html[index lifecycle management (ILM)] +* Support for {apm-server-ref-v}/ilm.html[index lifecycle management (ILM)] [[release-highlights-7.1.0]] === APM version 7.1.0 From 4825c30d03c680d928387da887e31db8fb08bb62 Mon Sep 17 00:00:00 2001 From: Brandon Morelli Date: Wed, 5 Jun 2019 08:13:04 -0700 Subject: [PATCH 5/5] docs: few more updates --- docs/ilm.asciidoc | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) diff --git a/docs/ilm.asciidoc b/docs/ilm.asciidoc index 0798c0866b5..bd51490aaf6 100644 --- a/docs/ilm.asciidoc +++ b/docs/ilm.asciidoc @@ -9,7 +9,7 @@ by automating rollovers to a new index when the existing index reaches a specifi You can view and edit the index lifecycle policies in the *Index lifecycle policies* UI in {kib}. For more information about working with the UI, -see {kibana-ref}/index-lifecycle-policies.html[Index lifecyle policies]. +see {kibana-ref}/index-lifecycle-policies.html[Index lifecycle policies]. APM Server currently offers two ways to get started with ILM: @@ -112,11 +112,16 @@ GET _ilm/policy <3> Rollover after _7 days_ <4> Priority for recovering your indices after a node restart. Higher priorities are recovered first. -Your indices are now configured to use index lifecylce management. Go ahead and <>. +Your indices are now configured to use index lifecycle management. Go ahead and <>. -// [float] -// [[ilm-upgrade]] -// ==== ILM default policy upgrades +NOTE: If you switch between ILM enabled/disabled multiple times, +you should set `setup.template.overwrite` to `true` to ensure a complete setup. + +[float] +==== ILM default policy upgrades + +If you decide to customize the default ILM policy, any customizations will be overwritten when you upgrade APM Server. +Default policies are also subject to change in future releases // Policies are versioned so they can change. // Indices are versioned so they can change.