docs: update apmschema url to use apm-data

specs/agents/metadata.md

@@ -192,7 +192,7 @@ Python:
 
 ### Cloud Provider Metadata
 
-[Cloud provider metadata](https://github.com/elastic/apm-server/blob/main/docs/spec/v2/metadata.json)
+[Cloud provider metadata](https://github.com/elastic/apm-data/blob/main/input/elasticapm/docs/spec/v2/metadata.json)
 is collected from local cloud provider metadata services:
 
 - availability_zone

specs/agents/metrics-health.md

@@ -27,7 +27,7 @@ If an agent supports these metrics, the metrics MUST be captured for all events
 Therefore an event MUST be taken into account for these metrics if an attempt is made to submit it to the internal reporting queue, regardless of the success.
 Events dropped before this attempt (e.g. due to [sampling](tracing-sampling.md) or [transaction_max_spans](handling-huge-traces/tracing-spans-limit.md)) are NOT counted.
 
-The `agent.events.total` metric MUST have the label `event_type`. The possible values for `event_type` are the lowercase names of the events according to the [spec](https://github.com/elastic/apm-server/tree/main/docs/spec/v2) (e.g. `transaction`, `metricset`) or the value `undefined`.
+The `agent.events.total` metric MUST have the label `event_type`. The possible values for `event_type` are the lowercase names of the events according to the [spec](https://github.com/elastic/apm-server/tree/main/input/elasticapm/docs/spec/v2) (e.g. `transaction`, `metricset`) or the value `undefined`.
 Agents SHOULD attempt to assign the appropriate label value based on the counted event. If this would impose significant implementation overhead, the value `undefined` MUST be used instead.
 
 The `agent.events.dropped` metric MUST have a value for the `reason` label. The following values MUST be used:

specs/agents/process-new-fields.md

@@ -11,7 +11,7 @@ it would create new fields in the index and then stop using it when we standardi
 For example `context.elasticsearch.url` in the intake API becomes `elasticsearch.url` in Elasticsearch, `context.elasticsearch.error_reason` becomes `elasticsearch.error_reason` etc.
 * The proposal needs to specify which fields should be indexed.
 An APM Server person might need to assist here to determine the right data type for the indexed fields.
-* The proposal should include the suggested [JSON Schema](https://github.com/elastic/apm-server/tree/main/docs/spec/v2) changes for all new fields.
+* The proposal should include the suggested [JSON Schema](https://github.com/elastic/apm-data/tree/main/input/elasticapm/docs/spec/v2) changes for all new fields.
 This forces alignment on the exact field names, JSON data type, length restrictions etc.
 * Make sure to check if [ECS](https://github.com/elastic/ecs) has defined appropriate fields for what you're proposing.
 * Agents should agree to the changes in a voting format (checkboxes),

specs/agents/tracing-instrumentation-aws-lambda.md

@@ -27,7 +27,7 @@ Field | Value | Description | Source
 `context.cloud.origin.*` | - | Do not set these fields if trigger type is `other`.  | Trigger specific.
 `context.service.origin.*` | - | Do not set these fields if trigger type is `other`. | Trigger specific.
 
-Note that `faas.*` fields *are not* nested under the context property [in the intake api](https://github.com/elastic/apm-server/blob/main/docs/spec/v2/transaction.json)! `faas` is a top-level key on the transaction.
+Note that `faas.*` fields *are not* nested under the context property [in the intake api](https://github.com/elastic/apm-server/blob/main/input/elasticapm/docs/spec/v2/transaction.json)! `faas` is a top-level key on the transaction.
 
 ### Overwriting Metadata
 Automatically capturing cloud metadata doesn't work reliably from a Lambda environment. Moreover, retrieving cloud metadata through an additional HTTP request may slowdown the lambda function / increase cold start behaviour. Therefore, the generic cloud metadata fetching should be disabled when the agent is running in a lambda context (for instance through checking for the existence of the `AWS_LAMBDA_FUNCTION_NAME` environment variable).