Course

This procedure is a part of a course that teaches you how to build a quickstart. If you haven't already, checkout the course introduction.

Each procedure in this course builds on top of the last one, so make sure you've completed the last procedure, send logs from your product before proceeding with this one.

Traces capture details of a single request as it moves through a system. They’re composed of spans, which are data structures that represent individual operations in the flow of execution.

New Relic, provides you a variety of ways to instrument your application to send traces to our Trace API.

In this lesson, you learn to send traces from your product using our telemetry software development kit (SDK).

Use our SDK

We offer an open source telemetry SDK in several of the most popular programming languages such as Python, Java, Node/TypeScript. These send data to our data ingest APIs, including our Trace API.

In this lesson, you learn how to install and use the Python telemetry SDK to report your first span to New Relic.

Report your first span

Step 1 of 7

Change to the send-traces/flashDB direcrory of the course repository.

bash

$cd ../../send-traces/flashDB

Step 2 of 7

If you haven't already, install the newrelic-telemetry-sdk package.

bash

$pip install newrelic-telemetry-sdk

Step 3 of 7

Open db.py file in the IDE of your choice and configure the SpanClient.

1
import os
2
import random
3
import datetime
4
from sys import getsizeof
5
import psutil
6

7
from newrelic_telemetry_sdk import MetricClient, GaugeMetric, CountMetric, SummaryMetric
8
from newrelic_telemetry_sdk import EventClient, Event
9
from newrelic_telemetry_sdk import LogClient, Log
10
from newrelic_telemetry_sdk import SpanClient
11

12
metric_client = MetricClient(os.environ["NEW_RELIC_LICENSE_KEY"])
13
event_client = EventClient(os.environ["NEW_RELIC_LICENSE_KEY"])
14
log_client = LogClient(os.environ["NEW_RELIC_LICENSE_KEY"])
15
span_client = SpanClient(os.environ["NEW_RELIC_LICENSE_KEY"])
16

17
db = {}
18
stats = {
19
    "read_response_times": [],
20
    "read_errors": 0,
21
    "read_count": 0,
22
    "create_response_times": [],
23
    "create_errors": 0,
24
    "create_count": 0,
25
    "update_response_times": [],
26
    "update_errors": 0,
27
    "update_count": 0,
28
    "delete_response_times": [],
29
    "delete_errors": 0,
30
    "delete_count": 0,
31
    "cache_hit": 0,
32
}
33
last_push = {
34
    "read": datetime.datetime.now(),
35
    "create": datetime.datetime.now(),
36
    "update": datetime.datetime.now(),
37
    "delete": datetime.datetime.now(),
38
}
39

40
def read(key):
41

42
    print(f"Reading...")
43

44
    if random.randint(0, 30) > 10:
45
        stats["cache_hit"] += 1
46

47
    stats["read_response_times"].append(random.uniform(0.5, 1.0))
48
    if random.choice([True, False]):
49
        stats["read_errors"] += 1
50
    stats["read_count"] += 1
51
    try_send("read")
52

53
def create(key, value):
54

55
    print(f"Writing...")
56

57
    db[key] = value
58
    stats["create_response_times"].append(random.uniform(0.5, 1.0))
59
    if random.choice([True, False]):
60
        stats["create_errors"] += 1
61
    stats["create_count"] += 1
62
    try_send("create")
63

64
def update(key, value):
65

66
    print(f"Updating...")
67

68
    db[key] = value
69
    stats["update_response_times"].append(random.uniform(0.5, 1.0))
70
    if random.choice([True, False]):
71
        stats["update_errors"] += 1
72
    stats["update_count"] += 1
73
    try_send("update")
74

75
def delete(key):
76

77
    print(f"Deleting...")
78

79
    db.pop(key, None)
80
    stats["delete_response_times"].append(random.uniform(0.5, 1.0))
81
    if random.choice([True, False]):
82
        stats["delete_errors"] += 1
83
    stats["delete_count"] += 1
84
    try_send("delete")
85

86
def try_send(type_):
87

88
    print("try_send")
89

90
    now = datetime.datetime.now()
91
    interval_ms = (now - last_push[type_]).total_seconds() * 1000
92
    if interval_ms >= 2000:
93
        send_metrics(type_, interval_ms)
94
        send_event(type_)
95
        send_logs()
96

97
def send_metrics(type_, interval_ms):
98
    
99
    print("sending metrics...")
100

101
    keys = GaugeMetric("fdb_keys", len(db))
102
    db_size = GaugeMetric("fdb_size", getsizeof(db))
103

104
    errors = CountMetric(
105
        name=f"fdb_{type_}_errors",
106
        value=stats[f"{type_}_errors"],
107
        interval_ms=interval_ms
108
    )
109

110
    cache_hits = CountMetric(
111
        name=f"fdb_cache_hits",
112
        value=stats["cache_hit"],
113
        interval_ms=interval_ms
114
    )
115

116
    response_times = stats[f"{type_}_response_times"]
117
    response_time_summary = SummaryMetric(
118
        f"fdb_{type_}_responses",
119
        count=len(response_times),
120
        min=min(response_times),
121
        max=max(response_times),
122
        sum=sum(response_times),
123
        interval_ms=interval_ms,
124
    )
125

126
    batch = [keys, db_size, errors, cache_hits, response_time_summary]
127
    response = metric_client.send_batch(batch)
128
    response.raise_for_status()
129
    print("Sent metrics successfully!")
130
    clear(type_)
131

132
def send_event(type_):
133

134
    print("sending event...")
135

136
    count = Event(
137
        "fdb_method", {"method": type_}
138
    )
139

140
    response = event_client.send_batch(count)
141
    response.raise_for_status()
142
    print("Event sent successfully!")
143

144
def send_logs():
145

146
    print("sending log...")
147

148
    process = psutil.Process(os.getpid())
149
    memory_usage = process.memory_percent()
150

151
    log = Log("FlashDB is using " + str(round(memory_usage * 100, 2)) + "% memory")
152

153
    response = log_client.send(log)
154
    response.raise_for_status()
155
    print("Log sent successfully!")
156

157
def clear(type_):
158
    stats[f"{type_}_response_times"] = []
159
    stats[f"{type_}_errors"] = 0
160
    stats["cache_hit"] = 0
161
    stats[f"{type_}_count"] = 0
162
    last_push[type_] = datetime.datetime.now()

db.py

Important

This example expects an environment variable called $NEW_RELIC_LICENSE_KEY.

Step 4 of 7

Instrument your app to report a span to New Relic.

1
import os
2
import random
3
import datetime
4
from sys import getsizeof
5
import psutil
6
import time
7

8
from newrelic_telemetry_sdk import MetricClient, GaugeMetric, CountMetric, SummaryMetric
9
from newrelic_telemetry_sdk import EventClient, Event
10
from newrelic_telemetry_sdk import LogClient, Log
11
from newrelic_telemetry_sdk import SpanClient, Span
12

13
metric_client = MetricClient(os.environ["NEW_RELIC_LICENSE_KEY"])
14
event_client = EventClient(os.environ["NEW_RELIC_LICENSE_KEY"])
15
log_client = LogClient(os.environ["NEW_RELIC_LICENSE_KEY"])
16
span_client = SpanClient(os.environ["NEW_RELIC_LICENSE_KEY"])
17

18
db = {}
19
stats = {
20
    "read_response_times": [],
21
    "read_errors": 0,
22
    "read_count": 0,
23
    "create_response_times": [],
24
    "create_errors": 0,
25
    "create_count": 0,
26
    "update_response_times": [],
27
    "update_errors": 0,
28
    "update_count": 0,
29
    "delete_response_times": [],
30
    "delete_errors": 0,
31
    "delete_count": 0,
32
    "cache_hit": 0,
33
}
34
last_push = {
35
    "read": datetime.datetime.now(),
36
    "create": datetime.datetime.now(),
37
    "update": datetime.datetime.now(),
38
    "delete": datetime.datetime.now(),
39
}
40

41
def read(key):
42

43
    print(f"Reading...")
44

45
    if random.randint(0, 30) > 10:
46
        stats["cache_hit"] += 1
47

48
    stats["read_response_times"].append(random.uniform(0.5, 1.0))
49
    if random.choice([True, False]):
50
        stats["read_errors"] += 1
51
    stats["read_count"] += 1
52
    try_send("read")
53

54
def create(key, value):
55

56
    print(f"Writing...")
57

58
    db[key] = value
59
    stats["create_response_times"].append(random.uniform(0.5, 1.0))
60
    if random.choice([True, False]):
61
        stats["create_errors"] += 1
62
    stats["create_count"] += 1
63
    try_send("create")
64

65
def update(key, value):
66

67
    print(f"Updating...")
68

69
    db[key] = value
70
    stats["update_response_times"].append(random.uniform(0.5, 1.0))
71
    if random.choice([True, False]):
72
        stats["update_errors"] += 1
73
    stats["update_count"] += 1
74
    try_send("update")
75

76
def delete(key):
77

78
    print(f"Deleting...")
79

80
    db.pop(key, None)
81
    stats["delete_response_times"].append(random.uniform(0.5, 1.0))
82
    if random.choice([True, False]):
83
        stats["delete_errors"] += 1
84
    stats["delete_count"] += 1
85
    try_send("delete")
86

87
def try_send(type_):
88

89
    print("try_send")
90

91
    now = datetime.datetime.now()
92
    interval_ms = (now - last_push[type_]).total_seconds() * 1000
93
    if interval_ms >= 2000:
94
        send_metrics(type_, interval_ms)
95
        send_event(type_)
96
        send_logs()
97

98
def send_metrics(type_, interval_ms):
99
    
100
    print("sending metrics...")
101

102
    keys = GaugeMetric("fdb_keys", len(db))
103
    db_size = GaugeMetric("fdb_size", getsizeof(db))
104

105
    errors = CountMetric(
106
        name=f"fdb_{type_}_errors",
107
        value=stats[f"{type_}_errors"],
108
        interval_ms=interval_ms
109
    )
110

111
    cache_hits = CountMetric(
112
        name=f"fdb_cache_hits",
113
        value=stats["cache_hit"],
114
        interval_ms=interval_ms
115
    )
116

117
    response_times = stats[f"{type_}_response_times"]
118
    response_time_summary = SummaryMetric(
119
        f"fdb_{type_}_responses",
120
        count=len(response_times),
121
        min=min(response_times),
122
        max=max(response_times),
123
        sum=sum(response_times),
124
        interval_ms=interval_ms,
125
    )
126

127
    batch = [keys, db_size, errors, cache_hits, response_time_summary]
128
    response = metric_client.send_batch(batch)
129
    response.raise_for_status()
130
    print("Sent metrics successfully!")
131
    clear(type_)
132

133
def send_event(type_):
134

135
    print("sending event...")
136

137
    count = Event(
138
        "fdb_method", {"method": type_}
139
    )
140

141
    response = event_client.send_batch(count)
142
    response.raise_for_status()
143
    print("Event sent successfully!")
144

145
def send_logs():
146

147
    print("sending log...")
148

149
    process = psutil.Process(os.getpid())
150
    memory_usage = process.memory_percent()
151

152
    log = Log("FlashDB is using " + str(round(memory_usage * 100, 2)) + "% memory")
153

154
    response = log_client.send(log)
155
    response.raise_for_status()
156
    print("Log sent successfully!")
157

158
def send_spans():
159

160
    print("sending span...")
161

162
    with Span(name="sleep") as span:
163
        time.sleep(0.5)
164

165
    response = span_client.send(span)
166
    response.raise_for_status()
167
    print("Span sleep sent successfully!")
168

169
def clear(type_):
170
    stats[f"{type_}_response_times"] = []
171
    stats[f"{type_}_errors"] = 0
172
    stats["cache_hit"] = 0
173
    stats[f"{type_}_count"] = 0
174
    last_push[type_] = datetime.datetime.now()

db.py

Here, you instrument your platform to send a simple sleep span to New Relic.

Step 5 of 7

Amend the try_send module to send the span every 2 second.

1
import os
2
import random
3
import datetime
4
from sys import getsizeof
5
import psutil
6
import time
7

8
from newrelic_telemetry_sdk import MetricClient, GaugeMetric, CountMetric, SummaryMetric
9
from newrelic_telemetry_sdk import EventClient, Event
10
from newrelic_telemetry_sdk import LogClient, Log
11
from newrelic_telemetry_sdk import SpanClient, Span
12

13
metric_client = MetricClient(os.environ["NEW_RELIC_LICENSE_KEY"])
14
event_client = EventClient(os.environ["NEW_RELIC_LICENSE_KEY"])
15
log_client = LogClient(os.environ["NEW_RELIC_LICENSE_KEY"])
16
span_client = SpanClient(os.environ["NEW_RELIC_LICENSE_KEY"])
17

18
db = {}
19
stats = {
20
    "read_response_times": [],
21
    "read_errors": 0,
22
    "read_count": 0,
23
    "create_response_times": [],
24
    "create_errors": 0,
25
    "create_count": 0,
26
    "update_response_times": [],
27
    "update_errors": 0,
28
    "update_count": 0,
29
    "delete_response_times": [],
30
    "delete_errors": 0,
31
    "delete_count": 0,
32
    "cache_hit": 0,
33
}
34
last_push = {
35
    "read": datetime.datetime.now(),
36
    "create": datetime.datetime.now(),
37
    "update": datetime.datetime.now(),
38
    "delete": datetime.datetime.now(),
39
}
40

41
def read(key):
42

43
    print(f"Reading...")
44

45
    if random.randint(0, 30) > 10:
46
        stats["cache_hit"] += 1
47

48
    stats["read_response_times"].append(random.uniform(0.5, 1.0))
49
    if random.choice([True, False]):
50
        stats["read_errors"] += 1
51
    stats["read_count"] += 1
52
    try_send("read")
53

54
def create(key, value):
55

56
    print(f"Writing...")
57

58
    db[key] = value
59
    stats["create_response_times"].append(random.uniform(0.5, 1.0))
60
    if random.choice([True, False]):
61
        stats["create_errors"] += 1
62
    stats["create_count"] += 1
63
    try_send("create")
64

65
def update(key, value):
66

67
    print(f"Updating...")
68

69
    db[key] = value
70
    stats["update_response_times"].append(random.uniform(0.5, 1.0))
71
    if random.choice([True, False]):
72
        stats["update_errors"] += 1
73
    stats["update_count"] += 1
74
    try_send("update")
75

76
def delete(key):
77

78
    print(f"Deleting...")
79

80
    db.pop(key, None)
81
    stats["delete_response_times"].append(random.uniform(0.5, 1.0))
82
    if random.choice([True, False]):
83
        stats["delete_errors"] += 1
84
    stats["delete_count"] += 1
85
    try_send("delete")
86

87
def try_send(type_):
88

89
    print("try_send")
90

91
    now = datetime.datetime.now()
92
    interval_ms = (now - last_push[type_]).total_seconds() * 1000
93
    if interval_ms >= 2000:
94
        send_metrics(type_, interval_ms)
95
        send_event(type_)
96
        send_logs()
97
        send_spans()
98

99
def send_metrics(type_, interval_ms):
100
    
101
    print("sending metrics...")
102

103
    keys = GaugeMetric("fdb_keys", len(db))
104
    db_size = GaugeMetric("fdb_size", getsizeof(db))
105

106
    errors = CountMetric(
107
        name=f"fdb_{type_}_errors",
108
        value=stats[f"{type_}_errors"],
109
        interval_ms=interval_ms
110
    )
111

112
    cache_hits = CountMetric(
113
        name=f"fdb_cache_hits",
114
        value=stats["cache_hit"],
115
        interval_ms=interval_ms
116
    )
117

118
    response_times = stats[f"{type_}_response_times"]
119
    response_time_summary = SummaryMetric(
120
        f"fdb_{type_}_responses",
121
        count=len(response_times),
122
        min=min(response_times),
123
        max=max(response_times),
124
        sum=sum(response_times),
125
        interval_ms=interval_ms,
126
    )
127

128
    batch = [keys, db_size, errors, cache_hits, response_time_summary]
129
    response = metric_client.send_batch(batch)
130
    response.raise_for_status()
131
    print("Sent metrics successfully!")
132
    clear(type_)
133

134
def send_event(type_):
135

136
    print("sending event...")
137

138
    count = Event(
139
        "fdb_method", {"method": type_}
140
    )
141

142
    response = event_client.send_batch(count)
143
    response.raise_for_status()
144
    print("Event sent successfully!")
145

146
def send_logs():
147

148
    print("sending log...")
149

150
    process = psutil.Process(os.getpid())
151
    memory_usage = process.memory_percent()
152

153
    log = Log("FlashDB is using " + str(round(memory_usage * 100, 2)) + "% memory")
154

155
    response = log_client.send(log)
156
    response.raise_for_status()
157
    print("Log sent successfully!")
158

159
def send_spans():
160

161
    print("sending span...")
162

163
    with Span(name="sleep") as span:
164
        time.sleep(0.5)
165

166
    response = span_client.send(span)
167
    response.raise_for_status()
168
    print("Span sleep sent successfully!")
169

170
def clear(type_):
171
    stats[f"{type_}_response_times"] = []
172
    stats[f"{type_}_errors"] = 0
173
    stats["cache_hit"] = 0
174
    stats[f"{type_}_count"] = 0
175
    last_push[type_] = datetime.datetime.now()

db.py

Your platform will now report this span every 2 seconds.

Step 6 of 7

Navigate to the root of your application at build-a-quickstart-lab/send-traces/flashDB.

Step 7 of 7

Run your services to verify that it is reporting the span.

bash

$python simulator.py
Writing...
try_send
Reading...
try_send
Reading...
try_send
Writing...
try_send
Writing...
try_send
Reading...
sending metrics...
Sent metrics successfully!
sending event...
Event sent successfully!
sending log...
Log sent successfully!
sending span...
Span sleep sent successfully!

Alternative Options

If the language SDK doesn’t fit your needs, try out one of our other options:

Existing Zipkin instrumentation: if you have an existing Zipkin implementation, you can simply change the endpoint to New Relic to report your data. Read our documentation to report data from existing Zipkin instrumentation.
Manual Implementation: If the previous options don’t fit your requirements, you can always manually instrument your own library to make a POST request to the New Relic Trace API.

Your platform is now reporting data to New Relic. Next, you observe this data in New Relic using dashboard.

Course

This procedure is a part of course that teaches you how to build a quickstart. Continue to next lesson, create a dashboard.

Send traces from your product

Course

Use our SDK

Report your first span

Important

Alternative Options

Course