Skip to content

Commit 8345078

Browse files
committed
LRS->xAPI
1 parent 9551e65 commit 8345078

File tree

1 file changed

+1
-1
lines changed

1 file changed

+1
-1
lines changed

doc/env_vars.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -111,7 +111,7 @@ The following options are used for advanced database performance tuning and may
111111
| `LRSQL_ENABLE_REACTIONS` | `enableReactions` | Whether or not to enable statement reactions. | `false` |
112112
| `LRSQL_REACTION_BUFFER_SIZE` | `reactionBufferSize` | Number of pending reactions to allow. Additional reactions will be dropped with a warning message. | `10000` |
113113
| `LRSQL_REACTION_VERSION` | `reactionVersion` | The xAPI version used when the reactor emits statements. Must be either `1.0.3` or `2.0.0`. | `1.0.3` |
114-
| `LRSQL_SUPPORTED_VERSIONS` | `supportedVersions` | A comma-separated list of supported LRS versions. This is used to validate incoming requests and ensure compatibility. | `1.0.3,2.0.0` |
114+
| `LRSQL_SUPPORTED_VERSIONS` | `supportedVersions` | A comma-separated list of supported xAPI versions. This is used to validate incoming requests and ensure compatibility. | `1.0.3,2.0.0` |
115115
| `LRSQL_ENABLE_STRICT_VERSION` | `enableStrictVersion` | Whether or not to enable strict versioning. When enabled, GET requests for statements using a 1.0.3 header will convert 2.0.0 statements by removing or normalizing fields. | `false` |
116116

117117
_NOTE:_ `LRSQL_STMT_RETRY_LIMIT` and `LRSQL_STMT_RETRY_BUDGET` are used to mitigate a rare scenario where specific Actors or Activities are updated many times in large concurrent batches. In this situation the DBMS can encounter locking and these settings are used to allow retries that eventually write all the conflicting transactions, but may incur performance degradation. If you are experiencing this situation the first step would be to look at why your data needs to rewrite specific Actors or Activities rapidly with different values, which could potentially solve it at the source. If the issue cannot be avoided by data design alone, another possible solution is reducing batch sizes to decrease or eliminate locks. As a last resort, increasing these settings will at least ensure the statements get written but as mentioned may incur a slowdown in concurrent throughput.

0 commit comments

Comments
 (0)