docs: clarify upgrade behavior (#12578)

Co-authored-by: Christian Haudum <[email protected]>

docs/sources/operations/scalability.md

@@ -17,7 +17,7 @@ and scaling for resource usage.
 The Query frontend has an in-memory queue that can be moved out into a separate process similar to the
 [Grafana Mimir query-scheduler](/docs/mimir/latest/operators-guide/architecture/components/query-scheduler/). This allows running multiple query frontends.
 
-To run with the Query Scheduler, the frontend needs to be passed the scheduler's address via `-frontend.scheduler-address` and the querier processes needs to be started with `-querier.scheduler-address` set to the same address. Both options can also be defined via the [configuration file]({{< relref "../configure/_index.md" >}}).
+To run with the Query Scheduler, the frontend needs to be passed the scheduler's address via `-frontend.scheduler-address` and the querier processes needs to be started with `-querier.scheduler-address` set to the same address. Both options can also be defined via the [configuration file](https://grafana.com/docs/loki/<LOKI_VERSION>/configure).
 
 It is not valid to start the querier with both a configured frontend and a scheduler address.
 

docs/sources/release-notes/v2-6.md

@@ -13,7 +13,7 @@ Grafana Labs is excited to announce the release of Loki 2.6. Here's a summary of
 - **Query multiple tenants at once.** We've introduced cross-tenant query federation, which allows you to issue one query to multiple tenants and get a single, consolidated result. This is great for scenarios where you need a global view of logs within your multi-tenant cluster. For more information on how to enable this feature, see [Multi-Tenancy]({{< relref "../operations/multi-tenancy.md" >}}). 
 - **Filter out and delete certain log lines from query results.** This is particularly useful in cases where users may accidentally write sensitive information to Loki that they do not want exposed. Users craft a LogQL query that selects the specific lines they're interested in, and then can choose to either filter out those lines from query results, or permanently delete them from Loki's storage. For more information, see [Logs Deletion]({{< relref "../operations/storage/logs-deletion.md" >}}).
 - **Improved query performance on instant queries.** Loki now splits instant queries with a large time range (for example, `sum(rate({app="foo"}[6h]))`) into several smaller sub-queries and executes them in parallel. Users don't need to take any action to enjoy this performance improvement; however, they can adjust the number of sub-queries generated by modifying the `split_queries_by_interval` configuration parameter, which currently defaults to `30m`. 
-- **Support Baidu AI Cloud as a storage backend.** Loki users can now use Baidu Object Storage (BOS) as their storage backend. See [bos_storage_config]({{< relref "../configure/_index.md#bos_storage_config" >}}) for details. 
+- **Support Baidu AI Cloud as a storage backend.** Loki users can now use Baidu Object Storage (BOS) as their storage backend. See [bos_storage_config](https://grafana.com/docs/loki/<LOKI_VERSION>/configure/) for details. 
 
 For a full list of all changes, look at the [CHANGELOG](https://github.com/grafana/loki/blob/main/CHANGELOG.md).
 
@@ -40,4 +40,4 @@ A summary of some of the more important fixes:
 - [PR 6152](https://github.com/grafana/loki/pull/6152) Fixed a scenario where live tailing of logs could cause unbounded ingester memory growth. 
 - [PR 5685](https://github.com/grafana/loki/pull/5685) Fixed a bug in Loki's push request parser that allowed users to send arbitrary non-string data as a log line. We now test that the pushed values are valid strings and return an error if values are not valid strings.
 - [PR 5799](https://github.com/grafana/loki/pull/5799) Fixed incorrect deduplication logic for cases where multiple log entries with the same timestamp exist. 
-- [PR 5888](https://github.com/grafana/loki/pull/5888) Fixed a bug in the [common configuration]({{< relref "../configure/_index.md#common" >}}) where the `instance_interface_names` setting was getting overwritten by the default ring configuration. 
+- [PR 5888](https://github.com/grafana/loki/pull/5888) Fixed a bug in the [common configuration](https://grafana.com/docs/loki/<LOKI_VERSION>/configure/#common) where the `instance_interface_names` setting was getting overwritten by the default ring configuration. 

docs/sources/setup/upgrade/_index.md

@@ -104,6 +104,10 @@ If no label is found matching the list, a value of `unknown_service` is applied.
 
 You can change this list by providing a list of labels to `discover_service_name` in the [limits_config](/docs/loki/<LOKI_VERSION>/configure/#limits_config) block.
 
+{{< admonition type="note" >}}
+If you are already using a `service_label`, Loki will not make a new assignment.
+{{< /admonition >}}
+
 **You can disable this by providing an empty value for `discover_service_name`.**
 
 #### Removed `shared_store` and `shared_store_key_prefix` from shipper configuration
@@ -171,7 +175,7 @@ The path prefix under which the delete requests are stored is decided by `-compa
 
 #### Configuration `async_cache_write_back_concurrency` and `async_cache_write_back_buffer_size` have been removed
 
-These configurations were redundant with the `Background` configuration in the [cache-config]({{< relref "../../configure#cache_config" >}}).
+These configurations were redundant with the `Background` configuration in the [cache-config](https://grafana.com/docs/loki/<LOKI_VERSION>/configure/#cache_config).
 
 `async_cache_write_back_concurrency` can be set with `writeback_goroutines`
 `async_cache_write_back_buffer_size` can be set with `writeback_buffer`
@@ -277,7 +281,7 @@ The TSDB index type has support for caching results for 'stats' and 'volume' que
 All of these are cached to the `results_cache` which is configured in the `query_range` config section.  By default, an in memory cache is used.
 
 #### Write dedupe cache is deprecated
-Write dedupe cache is deprecated because it not required by the newer single store indexes ([TSDB]({{< relref "../../operations/storage/tsdb" >}}) and [boltdb-shipper]({{< relref "../../operations/storage/boltdb-shipper" >}})).
+Write dedupe cache is deprecated because it not required by the newer single store indexes ([TSDB](https://grafana.com/docs/loki/<LOKI_VERSION>/operations/storage/tsdb/) and [boltdb-shipper](https://grafana.com/docs/loki/<LOKI_VERSION>/operations/storage/boltdb-shipper/)).
 If you using a [legacy index type](https://grafana.com/docs/loki/<LOKI_VERSION>/configure/storage/#index-storage), consider migrating to TSDB (recommended).
 
 #### Embedded cache metric changes
@@ -761,7 +765,7 @@ This histogram reports the distribution of log line sizes by file. It has 8 buck
 
 This creates a lot of series and we don't think this metric has enough value to offset the amount of series genereated so we are removing it.
 
-While this isn't a direct replacement, two metrics we find more useful are size and line counters configured via pipeline stages, an example of how to configure these metrics can be found in the [metrics pipeline stage docs]({{< relref "../../send-data/promtail/stages/metrics#counter" >}}).
+While this isn't a direct replacement, two metrics we find more useful are size and line counters configured via pipeline stages, an example of how to configure these metrics can be found in the [metrics pipeline stage docs](https://grafana.com/docs/loki/<LOKI_VERSION>/send-data/promtail/stages/metrics/#counter).
 
 #### `added Docker target` log message has been demoted from level=error to level=info
 
@@ -815,7 +819,7 @@ limits_config:
   retention_period: [30d]
 ```
 
-See the [retention docs]({{< relref "../../operations/storage/retention" >}}) for more info.
+See the [retention docs](https://grafana.com/docs/loki/<LOKI_VERSION>/operations/storage/retention/) for more info.
 
 #### Log messages on startup: proto: duplicate proto type registered:
 
@@ -1286,7 +1290,7 @@ If you happen to have `results_cache.max_freshness` set, use `limits_config.max_
 
 ### Promtail config removed
 
-The long deprecated `entry_parser` config in Promtail has been removed, use [pipeline_stages]({{< relref "../../send-data/promtail/configuration#pipeline_stages" >}}) instead.
+The long deprecated `entry_parser` config in Promtail has been removed, use [pipeline_stages](https://grafana.com/docs/loki/<LOKI_VERSION>/send-data/promtail/configuration/#pipeline_stages) instead.
 
 ### Upgrading schema to use boltdb-shipper and/or v11 schema
 
@@ -1616,7 +1620,7 @@ max_retries:
 
 Loki 1.4.0 vendors Cortex v0.7.0-rc.0 which contains [several breaking config changes](https://github.com/cortexproject/cortex/blob/v0.7.0-rc.0/CHANGELOG).
 
-In the [cache_config]({{< relref "../../configure#cache_config" >}}), `defaul_validity` has changed to `default_validity`.
+In the [cache_config](https://grafana.com/docs/loki/<LOKI_VERSION>/configure#cache_config), `defaul_validity` has changed to `default_validity`.
 
 If you configured your schema via arguments and not a config file, this is no longer supported. This is not something we had ever provided as an option via docs and is unlikely anyone is doing, but worth mentioning.