@cryptotaxi247 / netdata-1 / commits / 708efb41b

Support chart labels in alerts (#13290)

* chart labels for alerts * proper termination * use strchr * change if statement * change label variable. add docs * change doc * assign buf to temp * use new dictionary functions * reduce variable scope * reduce line length * make sure rrdcalc updates labels after inserted * reduce var scope * add rrdcalc.c for cmocka tests * Revert "add rrdcalc.c for cmocka tests" This reverts commit 5fe122adcf7abcbe6d67fa2ebd7c4ff8620cf9c8. * Fix cmocka unit tests * valgrind errors Co-authored-by: Vladimir Kobal <vlad@prokk.net>

Emmanuel Vasilakis committed Aug 16, 2022 at 10:33 UTC 708efb41bdf952c84b60326d40c07cc69e58d19c

15 files changed +147 -55

collectors/cgroups.plugin/tests/test_doubles.c

@@ -153,3 +153,7 @@ void sql_store_chart_label(uuid_t *chart_uuid, int source_type, char *label, cha

153 UNUSED(label);

154 UNUSED(value);

155 }

156 +

157 +void rrdcalc_update_rrdlabels(RRDSET *st) {

158 + (void)st;

159 +}

collectors/proc.plugin/proc_diskstats.c

@@ -852,6 +852,8 @@ static void add_labels_to_disk(struct disk *d, RRDSET *st) {

852 rrdlabels_add(st->state->chart_labels, "device_type", "virtual", RRDLABEL_SRC_AUTO);

853 break;

854 }

855 +

856 + rrdcalc_update_rrdlabels(st);

857 }

858

859 int do_proc_diskstats(int update_every, usec_t dt) {

collectors/proc.plugin/proc_mdstat.c

@@ -80,6 +80,7 @@ static inline void make_chart_obsolete(char *name, const char *id_modifier)

80 static void add_labels_to_mdstat(struct raid *raid, RRDSET *st) {

81 rrdlabels_add(st->state->chart_labels, "device", raid->name, RRDLABEL_SRC_AUTO);

82 rrdlabels_add(st->state->chart_labels, "raid_level", raid->level, RRDLABEL_SRC_AUTO);

83 + rrdcalc_update_rrdlabels(st);

84 }

86 int do_proc_mdstat(int update_every, usec_t dt)

collectors/proc.plugin/proc_net_wireless.c

@@ -200,6 +200,7 @@ static void configure_device(int do_status, int do_quality, int do_discarded_pac

200

201 static void add_labels_to_wireless(struct netwireless *w, RRDSET *st) {

202 rrdlabels_add(st->state->chart_labels, "device", w->name, RRDLABEL_SRC_AUTO);

203 + rrdcalc_update_rrdlabels(st);

204 }

205

206 int do_proc_net_wireless(int update_every, usec_t dt)

collectors/proc.plugin/sys_class_power_supply.c

@@ -114,6 +114,7 @@ void power_supply_free(struct power_supply *ps) {

114

115 static void add_labels_to_power_supply(struct power_supply *ps, RRDSET *st) {

116 rrdlabels_add(st->state->chart_labels, "device", ps->name, RRDLABEL_SRC_AUTO);

117 + rrdcalc_update_rrdlabels(st);

118 }

119

120 int do_sys_class_power_supply(int update_every, usec_t dt) {

collectors/proc.plugin/sys_fs_btrfs.c

@@ -451,6 +451,7 @@ static inline int find_all_btrfs_pools(const char *path) {

451 static void add_labels_to_btrfs(BTRFS_NODE *n, RRDSET *st) {

452 rrdlabels_add(st->state->chart_labels, "device", n->id, RRDLABEL_SRC_AUTO);

453 rrdlabels_add(st->state->chart_labels, "device_label", n->label, RRDLABEL_SRC_AUTO);

454 + rrdcalc_update_rrdlabels(st);

455 }

456

457 int do_sys_fs_btrfs(int update_every, usec_t dt) {

database/rrd.h

@@ -216,6 +216,7 @@ extern void rrdlabels_destroy(DICTIONARY *labels_dict);

216 extern void rrdlabels_add(DICTIONARY *dict, const char *name, const char *value, RRDLABEL_SRC ls);

217 extern void rrdlabels_add_pair(DICTIONARY *dict, const char *string, RRDLABEL_SRC ls);

218 extern void rrdlabels_get_value_to_buffer_or_null(DICTIONARY *labels, BUFFER *wb, const char *key, const char *quote, const char *null);

219 +extern void rrdlabels_get_value_to_char_or_null(DICTIONARY *labels, char **value, const char *key);

220

221 extern void rrdlabels_unmark_all(DICTIONARY *labels);

222 extern void rrdlabels_remove_all_unmarked(DICTIONARY *labels);

database/rrdcalc.c

+68 -1

@@ -35,6 +35,62 @@ inline const char *rrdcalc_status2string(RRDCALC_STATUS status) {

35 }

36 }

38 +char *rrdcalc_replace_variables(const char *line, RRDCALC *rc)

39 +{

40 + if (!line)

41 + return NULL;

42 +

43 + size_t pos = 0;

44 + char *temp = strdupz(line);

45 + char var[RRDCALC_VAR_MAX];

46 + char *m, *lbl_value = NULL;

47 +

48 + while ((m = strchr(temp + pos, '$'))) {

49 + int i=0;

50 + char *e = m;

51 + while (*e) {

52 + if (*e == ' ' || i == RRDCALC_VAR_MAX - 1) {

53 + break;

54 + }

55 + else

56 + var[i]=*e;

57 + e++;

58 + i++;

59 + }

60 + var[i]='\0';

61 + pos = m - temp + 1;

62 + if (!strcmp(var, RRDCALC_VAR_FAMILY)) {

63 + char *buf = find_and_replace(temp, var, (rc->rrdset && rc->rrdset->family) ? rc->rrdset->family : "", m);

64 + freez(temp);

65 + temp = buf;

66 + } else if (!strncmp(var, RRDCALC_VAR_LABEL, RRDCALC_VAR_LABEL_LEN)) {

67 + if(likely(rc->rrdset->state && rc->rrdset->state->chart_labels)) {

68 + rrdlabels_get_value_to_char_or_null(rc->rrdset->state->chart_labels, &lbl_value, var+RRDCALC_VAR_LABEL_LEN);

69 + if (lbl_value) {

70 + char *buf = find_and_replace(temp, var, lbl_value, m);

71 + freez(temp);

72 + temp = buf;

73 + freez(lbl_value);

74 + }

75 + }

76 + }

77 + }

78 +

79 + return temp;

80 +}

81 +

82 +void rrdcalc_update_rrdlabels(RRDSET *st) {

83 + RRDCALC *rc;

84 + for( rc = st->alarms; rc ; rc = rc->rrdset_next ) {

85 + if (rc->original_info) {

86 + if (rc->info)

87 + freez(rc->info);

88 +

89 + rc->info = rrdcalc_replace_variables(rc->original_info, rc);

90 + }

91 + }

92 +}

93 +

94 static void rrdsetcalc_link(RRDSET *st, RRDCALC *rc) {

95 RRDHOST *host = st->rrdhost;

@@ -83,6 +139,12 @@ static void rrdsetcalc_link(RRDSET *st, RRDCALC *rc) {

139

140 if(!rc->units) rc->units = strdupz(st->units);

141

142 + if (rc->original_info) {

143 + if (rc->info)

144 + freez(rc->info);

145 + rc->info = rrdcalc_replace_variables(rc->original_info, rc);

146 + }

147 +

148 time_t now = now_realtime_sec();

149 ALARM_ENTRY *ae = health_create_alarm_entry(

150 host,

@@ -427,7 +489,10 @@ inline RRDCALC *rrdcalc_create_from_template(RRDHOST *host, RRDCALCTEMPLATE *rt,

489 if(rt->recipient) rc->recipient = strdupz(rt->recipient);

490 if(rt->source) rc->source = strdupz(rt->source);

491 if(rt->units) rc->units = strdupz(rt->units);

430 - if(rt->info) rc->info = strdupz(rt->info);

492 + if(rt->info) {

493 + rc->info = strdupz(rt->info);

494 + rc->original_info = strdupz(rt->info);

495 + }

496

497 if (rt->classification) rc->classification = strdupz(rt->classification);

498 if (rt->component) rc->component = strdupz(rt->component);

@@ -543,6 +608,7 @@ inline RRDCALC *rrdcalc_create_from_rrdcalc(RRDCALC *rc, RRDHOST *host, const ch

608 if(rc->source) newrc->source = strdupz(rc->source);

609 if(rc->units) newrc->units = strdupz(rc->units);

610 if(rc->info) newrc->info = strdupz(rc->info);

611 + if(rc->original_info) newrc->original_info = strdupz(rc->original_info);

612

613 if (rc->classification) newrc->classification = strdupz(rc->classification);

614 if (rc->component) newrc->component = strdupz(rc->component);

@@ -586,6 +652,7 @@ void rrdcalc_free(RRDCALC *rc) {

652 freez(rc->source);

653 freez(rc->units);

654 freez(rc->info);

655 + freez(rc->original_info);

656 freez(rc->classification);

657 freez(rc->component);

658 freez(rc->type);

database/rrdcalc.h

@@ -58,6 +58,7 @@ struct rrdcalc {

59 char *source; // the source of this alarm

60 char *units; // the units of the alarm

61 + char *original_info; // the original info field before any variable replacement

62 char *info; // a short description of the alarm

64 int update_every; // update frequency for the alarm

@@ -210,6 +211,7 @@ extern RRDCALC *rrdcalc_create_from_rrdcalc(RRDCALC *rc, RRDHOST *host, const ch

211 extern void rrdcalc_add_to_host(RRDHOST *host, RRDCALC *rc);

212 extern void dimension_remove_pipe_comma(char *str);

213 extern char *alarm_name_with_dim(char *name, size_t namelen, const char *dim, size_t dimlen);

214 +extern void rrdcalc_update_rrdlabels(RRDSET *st);

215

216 extern void rrdcalc_labels_unlink();

217 extern void rrdcalc_labels_unlink_alarm_from_host(RRDHOST *host);

@@ -221,4 +223,9 @@ static inline int rrdcalc_isrepeating(RRDCALC *rc) {

223 return 0;

224 }

225

226 +#define RRDCALC_VAR_MAX 100

227 +#define RRDCALC_VAR_FAMILY "$family"

228 +#define RRDCALC_VAR_LABEL "$label:"

229 +#define RRDCALC_VAR_LABEL_LEN (sizeof(RRDCALC_VAR_LABEL)-1)

230 +

231 #endif //NETDATA_RRDCALC_H

database/rrdlabels.c

+14 -1

@@ -620,7 +620,7 @@ void rrdlabels_add_pair(DICTIONARY *dict, const char *string, RRDLABEL_SRC ls) {

620 }

621

622 // ----------------------------------------------------------------------------

623 -// rrdlabels_get_to_buffer_or_null()

623 +// rrdlabels_get_value_to_buffer_or_null()

624

625 void rrdlabels_get_value_to_buffer_or_null(DICTIONARY *labels, BUFFER *wb, const char *key, const char *quote, const char *null) {

626 DICTIONARY_ITEM *acquired_item = dictionary_get_and_acquire_item(labels, key);

@@ -634,6 +634,17 @@ void rrdlabels_get_value_to_buffer_or_null(DICTIONARY *labels, BUFFER *wb, const

634 dictionary_acquired_item_release(labels, acquired_item);

635 }

636

637 +// ----------------------------------------------------------------------------

638 +// rrdlabels_get_value_to_char_or_null()

639 +

640 +void rrdlabels_get_value_to_char_or_null(DICTIONARY *labels, char **value, const char *key) {

641 + DICTIONARY_ITEM *acquired_item = dictionary_get_and_acquire_item(labels, key);

642 + RRDLABEL *lb = dictionary_acquired_item_value(acquired_item);

643 +

644 + *value = (lb && lb->label_value) ? strdupz(string2str(lb->label_value)) : NULL;

645 +

646 + dictionary_acquired_item_release(labels, acquired_item);

647 +}

648

649 // ----------------------------------------------------------------------------

650 // rrdlabels_unmark_all()

@@ -939,6 +950,8 @@ void rrdset_update_rrdlabels(RRDSET *st, DICTIONARY *new_rrdlabels) {

950 if (new_rrdlabels)

951 rrdlabels_migrate_to_these(st->state->chart_labels, new_rrdlabels);

952

953 + rrdcalc_update_rrdlabels(st);

954 +

955 // TODO - we should also cleanup sqlite from old new_rrdlabels that have been removed

956 rrdlabels_walkthrough_read(st->state->chart_labels, chart_label_store_to_sql_callback, st);

957 }

exporting/tests/netdata_doubles.c

@@ -255,3 +255,7 @@ void sql_store_chart_label(uuid_t *chart_uuid, int source_type, char *label, cha

255 (void)label;

256 (void)value;

257 }

258 +

259 +void rrdcalc_update_rrdlabels(RRDSET *st) {

260 + (void)st;

261 +}

health/REFERENCE.md

+37 -1

@@ -536,12 +536,48 @@ See our [simple patterns docs](/libnetdata/simple_pattern/README.md) for more ex

536

537 #### Alarm line `info`

538

539 -The info field can contain a small piece of text describing the alarm or template. This will be rendered in notifications and UI elements whenever the specific alarm is in focus. An example for the `ram_available` alarm is:

539 +The info field can contain a small piece of text describing the alarm or template. This will be rendered in

540 +notifications and UI elements whenever the specific alarm is in focus. An example for the `ram_available` alarm is:

541

542 ```yaml

543 info: percentage of estimated amount of RAM available for userspace processes, without causing swapping

544 ```

545

546 +info fields can contain special variables in their text that will be replaced during run-time to provide more specific

547 +alert information. Current variables supported are:

548 +

549 +| variable | description |

550 +| ---------| ----------- |

551 +| $family | Will be replaced by the family instance for the alert (e.g. eth0) |

552 +| $label: | Followed by a chart label name, this will replace the variable with the chart label's value |

553 +

554 +For example, an info field like the following:

555 +

556 +```yaml

557 +info: average inbound utilization for the network interface $family over the last minute

558 +```

559 +

560 +Will be rendered on the alert acting on interface `eth0` as:

561 +

562 +```yaml

563 +info: average inbound utilization for the network interface eth0 over the last minute

564 +```

565 +

566 +An alert acting on a chart that has a chart label named e.g. `target`, with a value of `https://netdata.cloud/`,

567 +can be enriched as follows:

568 +

569 +```yaml

570 +info: average ratio of HTTP responses with unexpected status over the last 5 minutes for the site $label:target

571 +```

572 +

573 +Will become:

574 +

575 +```yaml

576 +info: average ratio of HTTP responses with unexpected status over the last 5 minutes for the site https://netdata.cloud/

577 +```

578 +

579 +> Please note that variable names are case sensitive.

580 +

581 ## Expressions

582

583 Netdata has an internal [infix expression parser](/libnetdata/eval). This parses expressions and creates an internal

health/health_config.c

@@ -942,6 +942,8 @@ static int health_readfile(const char *filename, void *data) {

942 }

943 rc->info = strdupz(value);

944 strip_quotes(rc->info);

945 + rc->original_info = strdupz(value);

946 + strip_quotes(rc->original_info);

947 }

948 else if(hash == hash_delay && !strcasecmp(key, HEALTH_DELAY_KEY)) {

949 alert_cfg->delay = strdupz(value);

health/health_json.c

+2 -35

@@ -96,22 +96,7 @@ void health_alarm_entry2json_nolock(BUFFER *wb, ALARM_ENTRY *ae, RRDHOST *host)

96 , (ae->flags & HEALTH_ENTRY_FLAG_SILENCED)?"true":"false"

97 );

99 - char *replaced_info = NULL;

100 - if (likely(ae->info)) {

101 - char *m = NULL;

102 - replaced_info = strdupz(ae->info);

103 - size_t pos = 0;

104 - while ((m = strstr(replaced_info + pos, "$family"))) {

105 - char *buf = NULL;

106 - pos = m - replaced_info;

107 - buf = find_and_replace(replaced_info, "$family", ae->family ? ae->family : "", m);

108 - freez(replaced_info);

109 - replaced_info = strdupz(buf);

110 - freez(buf);

111 - }

112 - }

113 -

114 - health_string2json(wb, "\t\t", "info", replaced_info?replaced_info:"", ",\n");

99 + health_string2json(wb, "\t\t", "info", ae->info ? ae->info : "", ",\n");

100

101 if(unlikely(ae->flags & HEALTH_ENTRY_FLAG_NO_CLEAR_NOTIFICATION)) {

102 buffer_strcat(wb, "\t\t\"no_clear_notification\": true,\n");

@@ -127,7 +112,6 @@ void health_alarm_entry2json_nolock(BUFFER *wb, ALARM_ENTRY *ae, RRDHOST *host)

112

113 buffer_strcat(wb, "\t}");

114

130 - freez(replaced_info);

115 freez(edit_command);

116 }

117

@@ -182,21 +166,6 @@ static inline void health_rrdcalc2json_nolock(RRDHOST *host, BUFFER *wb, RRDCALC

166 char value_string[100 + 1];

167 format_value_and_unit(value_string, 100, rc->value, rc->units, -1);

168

185 - char *replaced_info = NULL;

186 - if (likely(rc->info)) {

187 - char *m;

188 - replaced_info = strdupz(rc->info);

189 - size_t pos = 0;

190 - while ((m = strstr(replaced_info + pos, "$family"))) {

191 - char *buf = NULL;

192 - pos = m - replaced_info;

193 - buf = find_and_replace(replaced_info, "$family", (rc->rrdset && rc->rrdset->family) ? rc->rrdset->family : "", m);

194 - freez(replaced_info);

195 - replaced_info = strdupz(buf);

196 - freez(buf);

197 - }

198 - }

199 -

169 char hash_id[GUID_LEN + 1];

170 uuid_unparse_lower(rc->config_hash_id, hash_id);

171

@@ -250,7 +219,7 @@ static inline void health_rrdcalc2json_nolock(RRDHOST *host, BUFFER *wb, RRDCALC

219 , rc->recipient?rc->recipient:host->health_default_recipient

220 , rc->source

221 , rc->units?rc->units:""

253 - , replaced_info?replaced_info:""

222 + , rc->info?rc->info:""

223 , rrdcalc_status2string(rc->status)

224 , (unsigned long)rc->last_status_change

225 , (unsigned long)rc->last_updated

@@ -322,8 +291,6 @@ static inline void health_rrdcalc2json_nolock(RRDHOST *host, BUFFER *wb, RRDCALC

291 buffer_strcat(wb, "\n");

292

293 buffer_strcat(wb, "\t\t}");

325 -

326 - freez(replaced_info);

294 }

295

296 //void health_rrdcalctemplate2json_nolock(BUFFER *wb, RRDCALCTEMPLATE *rt) {

health/health_log.c

+2 -17

@@ -512,23 +512,8 @@ inline ALARM_ENTRY* health_create_alarm_entry(

512 ae->old_value_string = strdupz(format_value_and_unit(value_string, 100, ae->old_value, ae->units, -1));

513 ae->new_value_string = strdupz(format_value_and_unit(value_string, 100, ae->new_value, ae->units, -1));

514

515 - char *replaced_info = NULL;

516 - if (likely(info)) {

517 - char *m;

518 - replaced_info = strdupz(info);

519 - size_t pos = 0;

520 - while ((m = strstr(replaced_info + pos, "$family"))) {

521 - char *buf = NULL;

522 - pos = m - replaced_info;

523 - buf = find_and_replace(replaced_info, "$family", (ae->family) ? ae->family : "", m);

524 - freez(replaced_info);

525 - replaced_info = strdupz(buf);

526 - freez(buf);

527 - }

528 - }

529 -

530 - if(replaced_info) ae->info = strdupz(replaced_info);

531 - freez(replaced_info);

515 + if (info)

516 + ae->info = strdupz(info);

517

518 ae->old_status = old_status;

519 ae->new_status = new_status;