The v3.0.0 release of `google-cloud-logging` is focused on improving usability of the library,

particularly on newer serverless environments.

If you experience issues or have questions, please file an [issue](https://github.com/googleapis/python-logging/issues).

Prior to v3.0.0, there were three `Handler` classes used for the Python logging standard library integration:

- [`AppEngineHandler`](https://github.com/googleapis/python-logging/blob/v2.7.0/google/cloud/logging_v2/handlers/app_engine.py)

- [`ContainerEngineHandler`](https://github.com/googleapis/python-logging/blob/v2.7.0/google/cloud/logging_v2/handlers/container_engine.py)

- [`CloudLoggingHandler`](https://github.com/googleapis/python-logging/blob/v2.7.0/google/cloud/logging_v2/handlers/handlers.py)

Google Cloud has grown, and adding a new handler class for each new product does not scale.

Recently, we have changed to support two more generic `Handler` classes instead:

- [`CloudLoggingHandler`](https://github.com/googleapis/python-logging/blob/v2.7.0/google/cloud/logging_v2/handlers/handlers.py)

- sends logs over the network (using gRPC or HTTP API calls)

- replaces `AppEngineHandler`

- [`StructuredLogHandler`](https://github.com/googleapis/python-logging/blob/v2.7.0/google/cloud/logging_v2/handlers/structured_log.py)

- exports logs in JSON format through standard out, to be parsed by an agent

- replaces `ContainerEngineHandler`

As of v3.0.0, `AppEngineHandler` and `ContainerEngineHandler` have been marked as deprecated and will not be updated.

They may be removed from the library in a future update.

Logs emitted by the library must be associated with a [montored-resource type](https://cloud.google.com/monitoring/api/resources),

indicating the compute environment the log originated from. Previously, the logs would default to

["global"](https://cloud.google.com/monitoring/api/resources#tag_global) when left unspecified.

Going forward, the library will attempt to determine the monitored-resource automatically if not explicitly set.

You can now log JSON data using the Python `logging` standard library integration.

The library supports two different methods:

1. Using the `json_fields` `extra` argument:

import logging

data_dict = {"hello": "world"}

logging.info("message field", extra={"json_fields": data_dict})

2. Logging a JSON-parsable string:

import logging

import json

data_dict = {"hello": "world"}

logging.info(json.dumps(data_dict))

Previously, the Logger class had four methods for sending logs of different types:

logger.log_text("hello world")

logger.log_struct({"hello": "world"})

logger.log_proto(proto_message)

logger.log_empty()

In v3.0.0, the library adds a generic `log()` method that will attempt to infer and log any type:

logger.log("hello world")

In v3.0.0, the library will be more forgiving if inputs are given in a different format than expected

logger.log("hello world", severity="warning")

logger.log({"hello": "world", "severity":"warning"})

logger.log("hello world", resource={"type":"global", "labels":[]})

There was a crashing bug when attempting to read back logs from non-project resources:

- `organizations/[ORGANIZATION_ID]/logs/[LOG_ID]`

- `billingAccounts/[BILLING_ACCOUNT_ID]/logs/[LOG_ID]`

- `folders/[FOLDER_ID]/logs/[LOG_ID]`

The v3.0.0 update fixes this issue.

The library supports sending logs using two different network protocols: gRPC and HTTP. Previously, there was an \

inconsistency in the implementations, resulting in unexpected behaviour when in HTTP mode.

As part of these changes, we introduced a new `max_size` argument to `list_entries` calls, which can be used to determine

how many results should be returned:

from google.cloud import logging_v2

client = logging_v2.Client()

client.list_entries(max_size=5)

* `exclusion_path` -> `log_exclusion_path`