1
00:00:00,200 --> 00:00:01,650
‫So in this new demo we're going

2
00:00:01,650 --> 00:00:04,290
‫to create a workflow where we are going

3
00:00:04,290 --> 00:00:06,170
‫to handle some errors.

4
00:00:06,170 --> 00:00:08,060
‫And first we're going to create a function

5
00:00:08,060 --> 00:00:09,650
‫in Lambda that is going to throw an error.

6
00:00:09,650 --> 00:00:11,280
‫So let's create a function.

7
00:00:11,280 --> 00:00:13,930
‫We'll use a blueprint and in it,

8
00:00:13,930 --> 00:00:16,583
‫I'm going to type step function.

9
00:00:17,550 --> 00:00:19,750
‫Step, so let's clear the filter,

10
00:00:19,750 --> 00:00:22,740
‫step-functions.

11
00:00:22,740 --> 00:00:26,120
‫Okay, so we have step-functions-error as a blueprint

12
00:00:26,120 --> 00:00:28,980
‫to throw an error and then we can configure Step functions

13
00:00:28,980 --> 00:00:30,670
‫to handle the catch or retry from this error.

14
00:00:30,670 --> 00:00:32,940
‫So let's configure this blueprint

15
00:00:32,940 --> 00:00:35,930
‫and I'll call it MyLambdaFunctionThatFails.

16
00:00:37,930 --> 00:00:40,090
‫We're going to create a new role

17
00:00:40,090 --> 00:00:42,840
‫and here is the code itself.

18
00:00:42,840 --> 00:00:45,390
‫So it's going to look at it and create a CustomError.

19
00:00:45,390 --> 00:00:46,740
‫And then to say, "This is a custom error!"

20
00:00:46,740 --> 00:00:49,233
‫So this is a function that just only fails.

21
00:00:50,280 --> 00:00:51,630
‫Let's create this function.

22
00:00:52,930 --> 00:00:54,790
‫And the function is created now as we can see,

23
00:00:54,790 --> 00:00:57,300
‫if we test this function and pass in

24
00:00:57,300 --> 00:00:59,410
‫whatever test events and create it.

25
00:00:59,410 --> 00:01:01,193
‫So let's just call it Foobar

26
00:01:02,350 --> 00:01:05,440
‫and create this and test this function.

27
00:01:05,440 --> 00:01:07,040
‫It gives us an error.

28
00:01:07,040 --> 00:01:08,020
‫This is a CustomError.

29
00:01:08,020 --> 00:01:08,853
‫Here's the error message

30
00:01:08,853 --> 00:01:11,110
‫and here is the stack trace of the error.

31
00:01:11,110 --> 00:01:13,180
‫And we're going to handle this error directly

32
00:01:13,180 --> 00:01:15,150
‫from our Step functions.

33
00:01:15,150 --> 00:01:17,030
‫So let's create a new state machine

34
00:01:18,330 --> 00:01:20,420
‫and we're going to author it from scratch.

35
00:01:20,420 --> 00:01:24,950
‫And for the definition, again, let's go into our code.

36
00:01:24,950 --> 00:01:28,950
‫So we're going to 1-error-handling, state-machine.json.

37
00:01:28,950 --> 00:01:32,130
‫Copy this and paste it here.

38
00:01:32,130 --> 00:01:34,220
‫Refresh the graph and we have the ideas.

39
00:01:34,220 --> 00:01:36,220
‫So we're going to invoke our function

40
00:01:36,220 --> 00:01:38,940
‫and if it goes fine, it goes into the end states.

41
00:01:38,940 --> 00:01:41,570
‫And there are some different kinds of error.

42
00:01:41,570 --> 00:01:44,200
‫So CustomErrorFallback, ReservedTypeFallback,

43
00:01:44,200 --> 00:01:47,100
‫and CatchAllFallback, okay?

44
00:01:47,100 --> 00:01:49,870
‫So if we look at the code itself,

45
00:01:49,870 --> 00:01:52,300
‫it's going to say how to handle each error.

46
00:01:52,300 --> 00:01:55,260
‫So let's fix the error in right here first.

47
00:01:55,260 --> 00:01:59,170
‫We copied the ARN and paste it here.

48
00:01:59,170 --> 00:02:00,230
‫Great.

49
00:02:00,230 --> 00:02:02,570
‫So it's going to invoke our function.

50
00:02:02,570 --> 00:02:05,010
‫And then if the error is of type CustomError

51
00:02:05,010 --> 00:02:06,340
‫which is the case here.

52
00:02:06,340 --> 00:02:10,100
‫So as we can see here, we are throwing a CustomError.

53
00:02:10,100 --> 00:02:13,290
‫Then it's going to wait one second

54
00:02:13,290 --> 00:02:16,070
‫and do two attempts and have a BackofRate of two.

55
00:02:16,070 --> 00:02:18,320
‫So there's going to be some retry.

56
00:02:18,320 --> 00:02:21,860
‫If the task is failed, there's gonna be more

57
00:02:21,860 --> 00:02:24,000
‫interval seconds, more attempts, and more back of

58
00:02:24,000 --> 00:02:26,480
‫and all the other types of error is going to go into

59
00:02:26,480 --> 00:02:28,250
‫the catch all or the States.ALL

60
00:02:28,250 --> 00:02:30,550
‫and try some more time, okay?

61
00:02:30,550 --> 00:02:31,383
‫Then there's a catch.

62
00:02:31,383 --> 00:02:34,290
‫So once the retries are exhausted

63
00:02:34,290 --> 00:02:37,170
‫then it's going to go into the Catch phase.

64
00:02:37,170 --> 00:02:39,850
‫And if it's the CustomError

65
00:02:39,850 --> 00:02:41,600
‫then you go into the CustomErrorFallback.

66
00:02:41,600 --> 00:02:42,900
‫If it's a test failure, it goes into

67
00:02:42,900 --> 00:02:45,830
‫the ReservedTypeFallback and States.ALL

68
00:02:45,830 --> 00:02:47,730
‫it goes into the CatchAllFallback.

69
00:02:47,730 --> 00:02:52,140
‫And each fallback is going to be difference in here, okay?

70
00:02:52,140 --> 00:02:56,323
‫So let's click on next, MyStateMachineError.

71
00:02:57,770 --> 00:02:59,570
‫We're going to create a new role

72
00:02:59,570 --> 00:03:01,840
‫and create this state machine.

73
00:03:01,840 --> 00:03:05,010
‫So it is successfully created, let's start the execution

74
00:03:05,010 --> 00:03:07,230
‫and we'll open this into a new browser tab

75
00:03:09,120 --> 00:03:10,850
‫and we're going into our workflow.

76
00:03:10,850 --> 00:03:14,190
‫So we are invoking our function and this is in progress

77
00:03:14,190 --> 00:03:15,980
‫and we are gonna get some errors.

78
00:03:15,980 --> 00:03:17,780
‫So let's wait a little bit, but if we look

79
00:03:17,780 --> 00:03:21,510
‫at the event history, so the Lambda function was started,

80
00:03:21,510 --> 00:03:22,950
‫and then it failed, okay?

81
00:03:22,950 --> 00:03:25,530
‫It failed with this CustomError.

82
00:03:25,530 --> 00:03:28,280
‫And then it went into being scheduled

83
00:03:28,280 --> 00:03:30,530
‫because it was being retried, okay?

84
00:03:30,530 --> 00:03:32,700
‫So there's retry because the error message was captured.

85
00:03:32,700 --> 00:03:36,310
‫So it was started again, then failed, then scheduled again,

86
00:03:36,310 --> 00:03:38,960
‫then started again, then failed.

87
00:03:38,960 --> 00:03:40,020
‫And then the task executed

88
00:03:40,020 --> 00:03:42,410
‫because we have exhausted all our retries.

89
00:03:42,410 --> 00:03:45,600
‫So the task exited and then it got cut

90
00:03:45,600 --> 00:03:48,630
‫by the CustomErrorFallback, okay?

91
00:03:48,630 --> 00:03:50,320
‫And then the execution was finished.

92
00:03:50,320 --> 00:03:54,160
‫So if we go in here, as we can see this failed twice

93
00:03:54,160 --> 00:03:57,150
‫and three times, then it gets an error being caught

94
00:03:57,150 --> 00:04:00,180
‫and it went into the CustomErrorFallback and succeeded.

95
00:04:00,180 --> 00:04:02,540
‫Now, if we change our Lambda function a little bit

96
00:04:02,540 --> 00:04:06,770
‫and say we're going to do a NotCustomError

97
00:04:06,770 --> 00:04:10,150
‫which is an error that wasn't being cut by our code.

98
00:04:10,150 --> 00:04:14,923
‫This is a different custom error, which is uncut.

99
00:04:15,990 --> 00:04:18,100
‫So we're going to deploy this, excuse me,

100
00:04:18,100 --> 00:04:19,200
‫and then click on test.

101
00:04:19,200 --> 00:04:22,580
‫So this is not the same kind of error, okay?

102
00:04:22,580 --> 00:04:24,470
‫And we start our step-function again.

103
00:04:24,470 --> 00:04:27,090
‫So we're going to start a new execution

104
00:04:28,040 --> 00:04:30,883
‫and open this in a new tab and start the execution.

105
00:04:33,450 --> 00:04:37,440
‫As we can see it's running, but now the error we get

106
00:04:38,340 --> 00:04:41,010
‫is that we have a difference error.

107
00:04:41,010 --> 00:04:44,150
‫And so we expect it to go into a different route now.

108
00:04:44,150 --> 00:04:46,090
‫So let's wait a little bit

109
00:04:46,090 --> 00:04:49,393
‫for all the execution to go through.

110
00:04:50,280 --> 00:04:53,560
‫And so the execution has succeeded.

111
00:04:53,560 --> 00:04:54,820
‫And if we look at it now

112
00:04:54,820 --> 00:04:57,690
‫we went into the ReservedTypeFallback.

113
00:04:57,690 --> 00:04:59,530
‫And if you look at the event history, as we can see

114
00:04:59,530 --> 00:05:02,610
‫in terms of elapsed time, this first one failed.

115
00:05:02,610 --> 00:05:05,460
‫And then we waited 30 seconds to have

116
00:05:05,460 --> 00:05:07,820
‫the second one to be started.

117
00:05:07,820 --> 00:05:11,850
‫And then we waited 60 seconds, so it doubled the wait time

118
00:05:11,850 --> 00:05:14,000
‫to go and to try it a third time.

119
00:05:14,000 --> 00:05:16,060
‫And then it failed twice, so if we have three times,

120
00:05:16,060 --> 00:05:18,230
‫so it went into the TaskStateExited

121
00:05:18,230 --> 00:05:19,940
‫and then the execution finished.

122
00:05:19,940 --> 00:05:21,900
‫So it's really, really cool to see here

123
00:05:21,900 --> 00:05:24,140
‫that's based on the type of error, we have different

124
00:05:24,140 --> 00:05:26,280
‫retry logic and we have different catch logic.

125
00:05:26,280 --> 00:05:28,260
‫This retry logic was very, very quick,

126
00:05:28,260 --> 00:05:30,830
‫whereas the retry logic here was much longer,

127
00:05:30,830 --> 00:05:33,940
‫it was every 30 seconds and it went into a different branch.

128
00:05:33,940 --> 00:05:35,120
‫So hopefully that helps for you

129
00:05:35,120 --> 00:05:37,710
‫to understand error handling in Step functions.

130
00:05:37,710 --> 00:05:38,543
‫I hope you liked it.

131
00:05:38,543 --> 00:05:40,480
‫And to clean up everything, you should go

132
00:05:40,480 --> 00:05:43,140
‫into the Step Functions and delete your State machines

133
00:05:43,140 --> 00:05:45,410
‫but it won't cost you anything to have them around

134
00:05:45,410 --> 00:05:46,330
‫if you want to.

135
00:05:46,330 --> 00:05:47,163
‫All right, that's it.

136
00:05:47,163 --> 00:05:48,820
‫I will see you in the next lecture.