Service Bus: Track 2 Preview 1 API surface #7196

richardpark-msft · 2020-01-31T22:11:35Z

This is the current surface for preview.1.

Some of the more interesting bits:

We have some nice type inference happening on the lock mode passed to getReceiver/getSessionReceiver. The messages returned from that receiver will appropriately provide settlement methods (complete(), abandon(), etc..) or not (if receiveAndDelete was the lock mode).
There's a SubscriptionRuleManager you can spawn from the top level ServiceBusClient to manage rules.
getDeadLetterReceiver() is a top level method in ServiceBusClient

sdk/servicebus/service-bus/review/service-bus.api.md

…for arch review

Some things that are missing - some settings (iteration specifically) aren't yet plumbed thorugh so we can't configure timeouts or max wait times properly.

sdk/servicebus/service-bus/review/service-bus.api.md

* receiveBatch() * options to the main methods * peek (diagnostics.peek)

Expose autcomplete options

…ack2-reviewing

* Removing some of the temporary naming that was in place as we removed older classe

- Remove the need for QueueAuths or SubscriptionAuths

- rename iterateMessages to getMessageIterator

sdk/servicebus/service-bus/review/service-bus.api.md

ramya-rao-a · 2020-03-18T01:52:38Z

sdk/servicebus/service-bus/review/service-bus.api.md

-    get sessionId(): string | undefined;
-    get sessionLockedUntilUtc(): Date | undefined;
+    sessionId: string | undefined;
+    sessionLockedUntilUtc: Date | undefined;


The sessionId and sessionLockedUntilUtc previously were not settable by the user. Now they are. Why?

Good point. We should make them read-only.

sdk/servicebus/service-bus/review/service-bus.api.md

ramya-rao-a · 2020-03-28T04:23:53Z

service-bus.api.md

+        [key: string]: any;
+    }): Promise<void>;
+    complete(): Promise<void>;
+    deadLetter(options?: DeadLetterOptions & {


options -> propertiesToModify ?

ramya-rao-a · 2020-03-28T04:25:15Z

service-bus.api.md

+// @public
+export interface CreateBatchOptions extends OperationOptions {
+    maxSizeInBytes?: number;
+    retryOptions?: RetryOptions;


I do recall that this was added upon my insistence in another PR, but I believe the retry options shouldn't be here.
We do not need to have retryOptions customizable by the user in every method. At the bare minimum it should appear in the client constructor. As an extra, it may appear in the getSender/getReceiver methods.

ramya-rao-a · 2020-03-28T04:26:11Z

service-bus.api.md

+
+export { delay }
+
+export { Delivery }


Delivery was being exported before as it was part of the service bus message. Since we no longer expose this in the message, we don't need to export it

ramya-rao-a · 2020-03-28T04:29:09Z

service-bus.api.md

+// @public
+export interface Receiver<ReceivedMessageT> {
+    close(): Promise<void>;
+    diagnostics: {


Side thought: After recent customer interviews, I wonder if the word diagnostics may be more understood as diagnostics for the receiver and less as diagnostics for the queue/subscription i.e a user might expect to find methods or properties indicating whether the receiver/connection is open or closed or broken or in the process of reconnecting etc

ramya-rao-a · 2020-03-28T04:30:35Z

service-bus.api.md

+    isReceivingMessages(): boolean;
+    receiveBatch(maxMessages: number, options?: ReceiveBatchOptions): Promise<ReceivedMessageT[]>;
+    receiveDeferredMessage(sequenceNumber: Long, options?: OperationOptions): Promise<ReceivedMessageT | undefined>;
+    receiveDeferredMessages(sequenceNumbers: Long[], options?: OperationOptions): Promise<ReceivedMessageT[]>;


We have the single and multiple form for receiving deferred messages, but not for peek or receiveBatch. What are we using to draw the line between which features get the single & multiple variations and which get just the multiple one?

ramya-rao-a · 2020-03-28T04:32:45Z

service-bus.api.md

+    getDeadLetterReceiver(queueName: string, receiveMode: "receiveAndDelete"): Receiver<ReceivedMessage>;
+    getDeadLetterReceiver(topicName: string, subscriptionName: string, receiveMode: "peekLock"): Receiver<ReceivedMessageWithLock>;
+    getDeadLetterReceiver(topicName: string, subscriptionName: string, receiveMode: "receiveAndDelete"): Receiver<ReceivedMessage>;
+    getReceiver(queueName: string, receiveMode: "peekLock"): Receiver<ReceivedMessageWithLock>;


Is there a way that we can continue to have receive mode as an enum and still get the ReceivedMessageT fun? My concern with using string literals is opening up the chances of spelling mistakes

ramya-rao-a · 2020-03-28T04:34:30Z

service-bus.api.md

+    diagnostics: {
+        peek(maxMessageCount?: number): Promise<ReceivedMessage[]>;
+        peekBySequenceNumber(fromSequenceNumber: Long, maxMessageCount?: number): Promise<ReceivedMessage[]>;
+    };


The diagnostics here appear to be the same as in the Receiver. Since SessionReceiver is extending the Receiver, should this be appearing here at all?

ramya-rao-a · 2020-03-28T04:35:10Z

service-bus.api.md

+export interface SessionReceiverOptions {
+    maxSessionAutoRenewLockDurationInSeconds?: number;
+    sessionId: string | undefined;
+}


This interface isn't being used anywhere

ramya-rao-a · 2020-03-28T04:35:56Z

service-bus.api.md

+
+export { TokenCredential }
+
+export { TokenType }


Why is TokenType being exported?

…k2-reviewing

ramya-rao-a · 2020-04-03T21:09:25Z

service-bus.api.md

+export class ServiceBusClient {
+  constructor(connectionString: string, options?: ServiceBusClientOptions);
+  constructor(
+    hostName: string,


hostName -> fullyQualifiedNamespace

ramya-rao-a · 2020-04-03T21:24:23Z

service-bus.api.md

+  close(): Promise<void>;
+  getDeadLetterReceiver(
+    queueName: string,
+    receiveMode: "peekLock",


We should create an issue to discuss the need to make receive mode a mandatory parameter

ramya-rao-a · 2020-04-03T21:32:47Z

service-bus.api.md

+
+// @public
+export interface GetReceiverOptions {
+  retryOptions?: RetryOptions;


The theory that one would want to have retry options for a sender/receiver that is different from the ones set on the client is not a proved one. This should be something we add when we have user feedback for the same. Since this can be additional, we can always add it back.

ramya-rao-a · 2020-04-03T21:34:23Z

service-bus.api.md

+
+// @public
+export interface SessionReceiverOptions {
+  maxSessionAutoRenewLockDurationInSeconds?: number;


milli seconds and maybe maxAutoRenewLockDurationInMs or autoRenewLockDurationInMs

ramya-rao-a · 2020-04-03T21:34:43Z

service-bus.api.md

+export interface SessionReceiverOptions {
+  maxSessionAutoRenewLockDurationInSeconds?: number;
+  retryOptions?: RetryOptions;
+  sessionId: string | undefined;


Mark as optional parameter instead of | undefined?

ramya-rao-a · 2020-04-03T21:41:16Z

service-bus.api.md

+    options?: GetReceiverOptions
+  ): Receiver<ReceivedMessage>;
+  getSender(queueOrTopicName: string, options?: GetSenderOptions): Sender;
+  getSessionReceiver(


We have feedback for this to be async. We need issue to talk about getReceiver and getSender to be async as well

ramya-rao-a · 2020-04-03T21:45:58Z

service-bus.api.md

+  constructor(connectionString: string, options?: ServiceBusClientOptions);
+  constructor(
+    hostName: string,
+    tokenCredential: TokenCredential,


tokenCredential -> credential

ramya-rao-a · 2020-04-03T21:46:59Z

service-bus.api.md

+export interface ServiceBusMessage {
+  body: any;
+  contentType?: string;
+  correlationId?: string | number | Buffer;


Feedback from @bterlson is that this should support UInt8array as well if it supports Buffer

ramya-rao-a · 2020-04-03T22:01:04Z

service-bus.api.md

+export interface SessionReceiver<
+  ReceivedMessageT extends ReceivedMessage | ReceivedMessageWithLock
+> extends Receiver<ReceivedMessageT> {
+  diagnostics: {


diagnostics can give the impression that one can have diagnostic operations on the state of the receiver. If the intention here is to avoid confusion between peek() the method and peeklock the mode, then we can experiment with browse first

ramya-rao-a · 2020-04-03T22:08:52Z

service-bus.api.md

+import { AmqpMessage } from "@azure/core-amqp";
+import { delay } from "@azure/core-amqp";
+import { Delivery } from "rhea-promise";
+import Long from "long";


Need an issue to discuss the long and short of Long

ramya-rao-a · 2020-04-03T22:19:34Z

service-bus.api.md

+    peekBySequenceNumber(
+      fromSequenceNumber: Long,
+      maxMessageCount?: number
+    ): Promise<ReceivedMessage[]>;


We can move the count and the sequence number to options along with operation options

richardpark-msft requested review from ramya-rao-a, bterlson, JoshLove-msft, hemanttanwar and KieranBrantnerMagee January 31, 2020 22:11

JoshLove-msft reviewed Feb 3, 2020

View reviewed changes

richardpark-msft mentioned this pull request Feb 26, 2020

Board Review: ServiceBus (Python and JavaScript) Azure/azure-sdk#1046

Closed

richardpark-msft closed this Feb 26, 2020

richardpark-msft force-pushed the richardpark-sb-track2-reviewing branch from 438c52f to 04f0de2 Compare February 26, 2020 21:16

richardpark-msft added 3 commits February 26, 2020 13:59

No changes, just ran prettier.

43ae28c

iterator model needs work but everything else is pretty much correct …

5085719

…for arch review

Looks like we're good for the first preview review of the surface.

c899c1f

Some things that are missing - some settings (iteration specifically) aren't yet plumbed thorugh so we can't configure timeouts or max wait times properly.

richardpark-msft reopened this Feb 27, 2020

richardpark-msft commented Feb 27, 2020

View reviewed changes

sdk/servicebus/service-bus/review/service-bus.api.md Outdated Show resolved Hide resolved

richardpark-msft requested a review from ShivangiReja February 27, 2020 04:30

richardpark-msft changed the title ~~Track2 initial review with @bterlson~~ Service Bus: Track 2 Preview 1 API surface Feb 27, 2020

richardpark-msft requested review from HarshaNalluru, yunhaoling, chradek and conniey February 27, 2020 04:31

ShivangiReja reviewed Feb 27, 2020

View reviewed changes

sdk/servicebus/service-bus/review/service-bus.api.md Outdated Show resolved Hide resolved

sdk/servicebus/service-bus/review/service-bus.api.md Outdated Show resolved Hide resolved

sdk/servicebus/service-bus/review/service-bus.api.md Outdated Show resolved Hide resolved

richardpark-msft added 7 commits February 28, 2020 18:17

Just updated with:

06e0bab

* receiveBatch() * options to the main methods * peek (diagnostics.peek)

Removed the "uselessemptycontext"

14d5471

Expose autcomplete options

Merge remote-tracking branch 'upstream/master' into richardpark-sb-tr…

682363d

…ack2-reviewing

* streamMessages() renamed to subscribe() to match track 2 for EventHubs

245159f

* Removing some of the temporary naming that was in place as we removed older classe

Fix auths parameter name.

eafc9cc

- Moved to a parent client model (ServiceBusClient)

5b0eb96

- Remove the need for QueueAuths or SubscriptionAuths

- adding options so users can trace or add in abortsignals

035e9e4

- rename iterateMessages to getMessageIterator

richardpark-msft commented Mar 6, 2020

View reviewed changes

sdk/servicebus/service-bus/review/service-bus.api.md Outdated Show resolved Hide resolved

richardpark-msft commented Mar 6, 2020

View reviewed changes

sdk/servicebus/service-bus/review/service-bus.api.md Outdated Show resolved Hide resolved

updated from latest checkin with message settlement methods moved

d9e7315

ramya-rao-a reviewed Mar 18, 2020

View reviewed changes

updated with getDeadLetterReceiver() and SubscriptionRuleManager

c21985a

ramya-rao-a reviewed Mar 21, 2020

View reviewed changes

sdk/servicebus/service-bus/review/service-bus.api.md Outdated Show resolved Hide resolved

ramya-rao-a reviewed Mar 21, 2020

View reviewed changes

sdk/servicebus/service-bus/review/service-bus.api.md Outdated Show resolved Hide resolved

updated as of 6d5abc8 (2020-03-27)

5a186d1

ramya-rao-a reviewed Mar 28, 2020

View reviewed changes

service-bus.api.md Outdated

export { TokenCredential }

export { TokenType }

Copy link

Contributor

ramya-rao-a Mar 28, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why is TokenType being exported?

ramya-rao-a added 2 commits April 2, 2020 18:55

Merge remote-tracking branch 'origin/master' into richardpark-sb-trac…

2aee588

…k2-reviewing

Update the stand-alone api view

a8dd85b

ramya-rao-a reviewed Apr 3, 2020

View reviewed changes

ramya-rao-a added the Service Bus label Jun 22, 2020

ramya-rao-a closed this Sep 26, 2020

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Service Bus: Track 2 Preview 1 API surface #7196

Service Bus: Track 2 Preview 1 API surface #7196

richardpark-msft commented Jan 31, 2020 •

edited

Loading

ramya-rao-a Mar 18, 2020

HarshaNalluru Mar 18, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Mar 28, 2020

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020 •

edited

Loading

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020

ramya-rao-a Apr 3, 2020

Service Bus: Track 2 Preview 1 API surface #7196

Service Bus: Track 2 Preview 1 API surface #7196

Conversation

richardpark-msft commented Jan 31, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ramya-rao-a Apr 3, 2020 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

richardpark-msft commented Jan 31, 2020 •

edited

Loading

ramya-rao-a Apr 3, 2020 •

edited

Loading