[[configuring-default-internal-ops-timeouts]]
=== Configuring the Default Internal Operations Timeouts

While Watcher is active, it often accesses different indices in Elasticsearch.
These can be internal indices used for its ongoing operation (such as the `.watches`
index where all the watches are stored) or as part of a watch execution via the
<<input-search, `search` input>>, <<transform-search, `search` transform>> or the
<<actions-index, `index` actions>>.

To to ensure that Watcher's workflow doesn't hang on long running search or
indexing operations, these operations time out after a set period of time. You can
change the default timeouts in `elasticsearch.yml`. The timeouts you can configure
are shown in the following table.

[[default-internal-ops-timeouts]]
[options="header"]
|======
| Name                                              | Default  | Description
| `watcher.internal.ops.search.default_timeout`     | 30s      | The default timeout for all internal search operations.
| `watcher.internal.ops.index.default_timeout`      | 60s      | The default timeout for all internal index operations.
| `watcher.internal.ops.bulk.default_timeout`       | 120s     | The default timeout for all internal bulk operations.
|======